Category: Sci & Tech

  • KRA To Cut More From Kenya Meta Content Creators Payouts

    KRA To Cut More From Kenya Meta Content Creators Payouts

    In Summary Meta has informedily informed thousands of Kenyan content creators that a 5 per cent withholding tax will apply to all earnings from its platforms starting next year.

    The move comes as the Kenya Revenue Authority steps up enforcement on income from digital content monetisation, aligning with existing tax rules introduced in 2023. Creators will receive net payments after the deduction, with the tax reflected in monthly statements.

    Meta has notified Kenyan content creators earning from Facebook, Instagram and other platforms that it will start deducting 5 per cent withholding tax on all payouts from January 1, 2026, to comply with Kenya’s tax laws.

    In an email notice to affected users, the social media giant said recent changes in local regulations require companies to withhold and remit tax directly to the Kenya Revenue Authority on payments to Kenya-based creators.

    “Kenya tax law now requires businesses to deduct and remit withholding tax to KRA on payments made to creators located in Kenya,” the notice reads. “As a result, Meta will deduct 5 per cent withholding tax from all payments made to you.”

    The company added that the deduction will appear in monthly remittance advice, and creators will receive the net amount in their accounts. Payments processed from December 2025 onwards will reflect the new rule.

    The decision brings Meta in line with a provision in the Finance Act 2023 that subjects income from digital content monetisation to 5 per cent withholding tax for resident creators, with non-residents facing 20 per cent.

    Platforms such as YouTube and TikTok have implemented similar deductions where applicable, but Meta’s rollout marks a significant expansion as its monetisation features, including In-Stream Ads and Ads on Reels, gained traction in Kenya last year.

    The tax applies at source, meaning creators can claim credit for the withheld amount when filing annual income tax returns with KRA.

    It forms part of broader government efforts to capture revenue from the booming digital economy, where influencer marketing, video content and online advertising have created new income streams for young Kenyans.

    Industry players have welcomed the clarity but expressed concern over the timing, coming amid rising living costs.

    “This is not a new tax but enforcement of an existing one,” said one Nairobi-based creator who received the notice. “Many of us already factor in taxes, but seeing it deducted upfront will feel like a bigger cut.”

    KRA has in recent years intensified compliance in the creative sector, with President William Ruto previously highlighting the need to tax digital earnings to support national revenue targets.

    The authority’s ongoing rollout of tools to validate income declarations from January 2026 is expected to make under-reporting harder.

    Meta rolled out monetisation for Kenyan creators in August 2024, allowing earnings from original videos and Reels. Global Partnerships Lead for Africa, Moon Baz, described the features at launch as a boost for the local creative industry.

    The new withholding requirement affects thousands active on Meta’s family of apps, many of whom rely on M-Pesa for payouts following a 2024 agreement facilitated by the government.

    Creators are advised to update their tax details on Meta’s platforms and consult KRA for guidance on claiming credits during annual filings. The deducted tax is creditable, not additional, for those who declare full income.

  • Google Unveils Gemini 3, Its Most Advanced AI Yet, With Immediate Integration Into Search

    Google Unveils Gemini 3, Its Most Advanced AI Yet, With Immediate Integration Into Search

    Google has intensified the global race for artificial intelligence dominance with the launch of Gemini 3, its most advanced model so far.

    The tech giant rolled out the system in San Francisco, describing it as a major leap in both intelligence and commercial readiness, and for the first time deployed it directly into Google Search from the moment of release.

    The instant integration marks a significant shift for Google, which previously took weeks or months to embed new AI models into widely used products.

    CEO Sundar Pichai called Gemini 3 “our most intelligent model” and framed the launch as a turning point in Google’s strategy to accelerate delivery and monetization across its ecosystem.

    Gemini 3 arrives at a moment when the industry is moving away from bragging rights on academic benchmarks and toward tools that can generate real revenue.

    Although Google highlighted that Gemini 3 leads several popular performance rankings, executives emphasized that the model is already powering money-making products, which is what investors are now watching most closely.

    Koray Kavukcuoglu, Google’s chief AI architect, said the company has shifted into a faster rhythm of both model development and distribution.

    He noted that the priority now is to get these capabilities into users’ hands as quickly as possible, reflecting Google’s urgency in a market where OpenAI and Anthropic are pushing out competing systems at rapid speed.

    Paying subscribers on Google’s premium AI plan gained immediate access to the new system through AI Mode in Search.

    The feature replaces the familiar list of web links with fully generated answers for complex queries, a format that once again raises concerns among publishers already grappling with declining traffic.

    Google has also introduced Gemini Agent, a new capability that can handle multi-step tasks such as sorting email inboxes, planning travel or managing ongoing digital chores without constant user prompts.

    The rollout brings Google closer to its long-term ambition of creating a universal assistant, an internal concept known as AlphaAssist that has been under development within DeepMind.

    In a demonstration, Google showed how Gemini can now generate interactive, visually rich responses through the redesigned Gemini app.

    A request to create a Van Gogh gallery produced an interface with detailed descriptions and visuals resembling a dedicated website.

    The change is likely to reignite debate about Google’s impact on the open web, since such immersive answers may reduce the need for users to visit other platforms.

    For enterprise customers, Google previewed Antigravity, an autonomous software development platform where AI agents can coordinate, plan and execute coding tasks.

    The unveiling signals Google’s intent to push deeper into AI-driven engineering tools, an area where rivals have been racing to secure early advantage.

    Gemini 3 launches in an environment of both excitement and skepticism.

    Analysts have warned that the rapid pace of AI innovation, soaring infrastructure costs and intense competition could be signs of an overheated market.

    The failure of notable model updates earlier in the year showed how quickly investor and public sentiment can turn when expectations are not met.

    Google is betting that speed, scale and deep integration will keep it ahead in the AI wars.

    With Gemini 3 now embedded across Search, the Gemini app and upcoming enterprise platforms, the company is challenging its rivals with a message that intelligence alone is no longer enough.

    What matters is getting advanced AI into billions of hands as quickly as possible, and on that front Google believes it has reclaimed momentum.

  • EXPOSED: How Hackers Exploited Years of Ignored Warnings to Bring State House to Its Knees

    EXPOSED: How Hackers Exploited Years of Ignored Warnings to Bring State House to Its Knees

    Government’s digital fortress crumbles as cybercriminals waltz through vulnerabilities flagged for years

    The warning signs were everywhere. For years, they had been flashing red like emergency beacons in the night, screaming for attention from the corridors of power. But in the plush offices of State House and across government ministries, those alarms were systematically ignored, the reports gathering dust while Kenya’s digital infrastructure stood naked before the world’s most ruthless hackers.

    On Monday morning, the chickens came home to roost with a vengeance that sent shockwaves through the entire nation. When Kenyans tried logging onto government websites, from the presidency’s own portal to critical ministries handling everything from health to immigration, they were met with a chilling sight: white supremacist messages blazing across their screens, the digital equivalent of graffiti spray-painted across the face of the nation.

    “Access denied by PCP.” “We will rise again.” “White power worldwide.” And most disturbingly, “14:88 Heil Hitler.” The messages were not just an attack on Kenya’s digital infrastructure but a brazen assault on the nation’s dignity, executed with the surgical precision that only comes from exploiting weaknesses that have been laid bare for anyone willing to look.

    The shadowy group calling itself PCP@Kenya didn’t just stumble upon these vulnerabilities by accident. They had a roadmap, meticulously drawn up year after year by Auditor General Nancy Gathungu, whose office had been sounding the alarm about catastrophically weak cybersecurity systems across government that were practically begging for exactly this kind of assault. 

    The scale of Monday’s breach was staggering. State House itself went dark. The ministries of Health, Education, Labour, Environment, ICT, Tourism, Energy, Water, and Interior all fell like dominoes. The Directorate of Criminal Investigations, supposedly one of the country’s premier law enforcement agencies, couldn’t protect its own digital doors. The Immigration Department, holding sensitive data on every person who enters and exits Kenya, was compromised. Even the Hustler Fund, President William Ruto’s signature economic programme, was rendered inaccessible.

    For hours, the digital face of the Kenyan government displayed the hateful insignia of extremists who had waltzed through security systems that, according to official reports, were about as sturdy as tissue paper in a rainstorm.

    It turns out the hackers had plenty of help, not from insiders, but from the government’s own spectacular negligence. The Auditor General’s reports for the financial year ending June 2023 had warned in stark terms that eCitizen, the government’s flagship online service platform handling over 5,000 services, was operating without an ICT policy, without a steering committee, without an approved business continuity plan, and without even a secondary backup site. 

    Read that again. The platform holding mountains of sensitive data on millions of Kenyans was running on digital infrastructure that would embarrass a small town internet café. It was a disaster waiting to happen, and happen it did. Days after that financial year ended, eCitizen was hammered by hackers, bringing government services to a grinding halt.

    In her 2024 report, Gathungu revealed that 39 National Government Constituencies Development Funds were operating without ICT policies. Thirteen water companies had implemented weak ICT policies and controls that were wide open to attack.  The pattern was clear and catastrophic: across the sprawling apparatus of the Kenyan state, cybersecurity was treated as an afterthought, a box to be ticked rather than a critical line of defence in an increasingly hostile digital world.

    The Communications Authority had also joined the chorus of doom, reporting between July and September alone, a stunning 842 cyber threat events targeting Kenyan systems. Their diagnosis was damning: inadequate system patching, limited user awareness, and the failure to keep up with AI-driven attacks were leaving government institutions sitting ducks for sophisticated cybercriminals. 

    The warnings weren’t just coming from local watchdogs. A Central Bank report revealed that Kenyan banks haemorrhaged Sh1.59 billion to hackers in 2024 alone, with attacks more than doubling from 173 in 2023 to 353 in 2024.  If the private sector was bleeding this badly despite having profit motives to secure their systems, what hope did cash-strapped government ministries have?

    The answer arrived on Monday morning in the form of racist taunts splashed across official government websites.

    But here’s where the story gets even more damning. This wasn’t Kenya’s first rodeo with cyber catastrophe. In July 2023, eCitizen was paralysed by hackers who claimed to be a Sudanese group protesting Kenya’s alleged interference in Sudan’s affairs. Then ICT Cabinet Secretary Eliud Owalo had assured Kenyans no data was lost. The government promised to do better. Clearly, those promises evaporated faster than morning dew in the Turkana sun.

    The pattern is sickeningly familiar to anyone who has watched Kenya’s government in action: identify a problem, commission a report, hold a press conference, file the report away, repeat. Since 2018, when Edward Ouko was Auditor General, the office has issued warning after warning about cybersecurity vulnerabilities. Many were ignored completely. Others were implemented half-heartedly, with about as much commitment as a New Year’s resolution by mid-January. 

    Even when cybersecurity measures were put in place, they were often cosmetic. A policy document here, a committee meeting there, but the fundamental weaknesses remained, like termites eating away at the foundations of a house while the owners admired the fresh coat of paint.

    The year before the latest attack, someone at the Ministry of Health had managed to override controls in the Integrated Finance Management and Information system, creating a phantom account used to loot an undisclosed amount of taxpayers’ money.  If internal actors could penetrate the systems that easily, what chance did Kenya have against sophisticated international hacking syndicates?

    Interior Principal Secretary Raymond Omollo, in announcing the breach, pointed fingers at PCP@Kenya and promised that those found culpable would “face the full force of the law.”  But here’s the uncomfortable question: who will face the full force of the law for the years of negligence that made this attack possible? Who will be held accountable for filing away report after report warning of exactly this scenario?

    The government has now activated multi-agency response teams, enhanced monitoring, and is working with private sector partners to strengthen cybersecurity. It’s the equivalent of installing burglar alarms after your house has been cleaned out, then standing in the empty living room assuring everyone you’re committed to home security.

    The hackers, whoever they really are behind the PCP@Kenya moniker, have done Kenya an inadvertent favour. They’ve exposed, in the most humiliating way possible, what happens when a government treats cybersecurity like an inconvenient suggestion rather than a critical national priority. They’ve demonstrated that in the digital age, weak passwords and missing backup systems are as dangerous as leaving the country’s borders unguarded.

    Kenya has invested billions in digitising government services, touting its commitment to a digital economy and the Fourth Industrial Revolution. But building digital highways without basic security is like constructing glass houses in a neighbourhood of stone-throwers. The vision is worthless without the foundation.

    The government insists that critical systems like eCitizen, NTSA, the Judiciary, KNEC, and the National Police Service were unaffected. Defence and Treasury also dodged the bullet.  That’s cold comfort when the presidency’s own website, along with a dozen major ministries, were defaced with Nazi propaganda for the world to see.

    As investigators now scramble to identify the perpetrators and understand their motives, the real question hanging over Kenya is this: will this latest humiliation finally force the government to take cybersecurity seriously, or will the response follow the familiar pattern of outrage, promises, and eventual complacency until the next attack arrives?

    The attack violated the Computer Misuse and Cybercrimes Act, the Kenya Information and Communications Act, and the Data Protection Act.  But those laws are useless if the systems they’re meant to protect are built on quicksand. You can’t legislate security into existence. You have to invest in it, prioritise it, and actually implement the recommendations that experts have been making for years.

    The hackers have spoken. They’ve demonstrated in the most public way possible that Kenya’s digital emperor has no clothes. The question now is whether the government will finally listen to the warnings it’s been receiving all along, or whether we’ll be writing this same story again in another year or two, with a different hacking group and the same preventable vulnerabilities.

    For now, websites are being restored and officials are assuring Kenyans that everything is under control. But the damage has been done, not just to government servers but to Kenya’s reputation as a technology hub and safe digital destination. In the ruthless world of cybersecurity, there are two types of organisations: those who have been hacked and know it, and those who have been hacked and don’t know it yet.

    Kenya now knows. The only question is what it will do with that knowledge.

  • CATASTROPHIC SECURITY BREACH: 4.8 MILLION KENYANS’ PRIVATE MEDICAL RECORDS PLUNDERED IN MASSIVE M-TIBA CYBERATTACK

    CATASTROPHIC SECURITY BREACH: 4.8 MILLION KENYANS’ PRIVATE MEDICAL RECORDS PLUNDERED IN MASSIVE M-TIBA CYBERATTACK

    Dark web criminals expose intimate diagnoses, ID numbers, insurance details in what could be Kenya’s worst-ever health data disaster

    Kenya is reeling from what security experts are calling a “catastrophic privacy violation” after hackers claimed to have stolen millions of medical and personal records from M-Tiba a digital health wallet used by millions of Kenyans .

    A cybercrime syndicate calling itself Kazu claims to have seized more than 17 million files totaling approximately 2.15 terabytes of data from M-Tiba’s servers , creating a chilling treasure trove of the nation’s most intimate health secrets now circulating in the criminal underworld.

    The scale of the breach is staggering.

    The stolen database allegedly contains sensitive information on 4.8 million Kenyan users  , exposing not just names and national ID numbers but deeply personal medical diagnoses, billing information, and treatment histories that patients never imagined would end up for sale on dark web forums.

    A SINISTER SAMPLE REVEALS THE HORROR

    The hackers brazenly shared a 2GB sample of their stolen treasure on their Telegram channel, containing what appear to be patients’ names, national ID numbers, dates of birth, phone contacts, and in some cases their medical diagnoses and billing information.

    This sample alone has already compromised the data of about 114,000 users, both account holders and their beneficiaries .

    But the nightmare extends far beyond individual patients

    The leaked files contain records from about 700 health facilities, with some scans showing full billing sheets and patient diagnostic summaries, including the names of doctors and insurance companies.

    In one set of documents, patient IDs, contact details, and treatment costs were listed alongside handwritten notes from medical staff .

    Imagine waking up to discover your HIV diagnosis, mental health treatment, or cancer records are now available to identity thieves, blackmailers, and scam artists prowling the internet’s darkest corners.

    For millions of Kenyans, this nightmare is now reality.

    DEAFENING SILENCE FROM THE GUARDIANS

    Despite the severity of the breach, M-Tiba’s operator CarePay has neither confirmed nor denied the breach, instead asking journalists to share copies of the leaked files to assist with their review.

    This tepid response has sparked fury among cybersecurity experts who note that precious hours are being wasted while criminals exploit the stolen data.

    “At M-TIBA, we take all matters of data security with the utmost seriousness,” CarePay claimed in what critics are calling an empty platitude devoid of concrete action or accountability.

    Adding to the scandal, the Office of the Data Protection Commissioner acknowledged awareness of the incident but declined to elaborate, citing they were not authorized to comment on an active matter.

    This bureaucratic stonewalling has left millions of vulnerable Kenyans in the dark about what has happened to their most sensitive information.

    The timing could not be more damning.

    Just two months ago, in August 2025, M-Tiba proudly announced it had received ISO/IEC 27001:2022 certification for its Information Security Management System.

    That international security badge now looks like cruel irony as the company grapples with one of Kenya’s most devastating data breaches.

    KENYA’S DIGITAL DREAMS BECOME CYBERSECURITY NIGHTMARES

    This catastrophe arrives as Kenya’s digital transformation accelerates at breakneck speed, often outpacing the security infrastructure needed to protect it.

    The Communications Authority recorded over 4.6 billion cyberattacks between April and June 2025, an 80% rise compared to the previous quarter, with most incidents involving phishing, ransomware, and data breaches targeting banks, telecommunications companies, and government systems .

    M-Tiba has been one of Kenya’s digital success stories since its 2016 launch through a partnership between CarePay, Safaricom, and the PharmAccess Foundation.

    The platform allows users to save and spend money specifically for healthcare and is used to distribute insurance benefits and government health subsidies.

    By 2024, the platform had attracted over 4 million users and partnerships with more than 3,000 hospitals across the country.

    But that very success made it an irresistible target for cybercriminals.

    The more Kenyans trusted M-Tiba with their health information, the more valuable that data became on the black market.

    THE KAZU SHADOW SYNDICATE

    The threat actor Kazu has recently emerged as a notably active group engaged in data leak activities, with credible sources tying the group to multiple security breaches involving unauthorized system access and attempts to sell stolen data on dark web marketplaces.

    The hackers are advertising the stolen M-Tiba database on the cybercrime forum [darkforums.st](http://darkforums.st) , where criminals trade in human misery.

    The group’s methods remain mysterious. They have not explained how they penetrated M-Tiba’s supposedly secure systems or when the intrusion occurred.

    What is clear is that they now possess an unprecedented window into the medical lives of millions of Kenyans.

    A LEGAL AND MORAL RECKONING AWAITS

    If confirmed, the M-Tiba breach would mark one of the most serious exposures of medical data since Kenya’s Data Protection Act came into force in 2019. The law classifies health records as sensitive personal information, requiring strict safeguards .

    Under Kenya’s Data Protection Act, a company is required to notify the ODPC of a personal data breach within 72 hours of becoming aware of it.

    Whether M-Tiba has complied with this legal obligation remains unclear, adding potential regulatory violations to the company’s mounting crisis.

    The consequences for victims are severe and permanent. Medical records represent some of the most sensitive personal information imaginable.

    Combined with national ID numbers and contact details, this leaked data creates a perfect storm for identity theft, insurance fraud, medical blackmail, and targeted scams against some of Kenya’s most vulnerable citizens.

    For pregnant women whose conditions are now exposed, for HIV patients whose status has been compromised, for mental health patients whose private struggles are now public, and for cancer patients whose diagnoses are being traded like commodities, this breach represents a violation that no apology can undo.

    As Kenya races toward a digital future, the M-Tiba catastrophe stands as a brutal reminder that without robust cybersecurity, every convenience comes with existential risk.

    The question now is not whether this will happen again, but how many more millions of Kenyans will have their privacy shattered before the country’s digital guardians take security seriously.

    The data is already out there.

    For 4.8 million Kenyans, it is too late to close the barn door. The damage is done, and the criminals are counting their profits.

  • Musk Launches Grokipedia To Rival ‘Left-biased’ Wikipedia

    Musk Launches Grokipedia To Rival ‘Left-biased’ Wikipedia

    After months of delays to “purge out the propaganda”, Elon Musk on Monday announced the launch of Grokipedia to rival online encyclopedia Wikipedia, which he has accused of ideological bias.

    The content of Grokipedia is generated by artificial intelligence (AI) and the generative AI assistant Grok.

    Elon Musk’s company xAI launched Grokipedia on Monday to compete with online encyclopedia Wikipedia, which he has accused of ideological bias.

    The site dubbed version 0.1 had more than 885,000 articles by Monday evening, compared to Wikipedia’s more than seven million in English.

    The launch came with the promise of a newer version 1.0, which Musk said would be “10X better” than the current live site, which he claimed is already “better than Wikipedia.”

    “The goal of Grok and Grokipedia.com is the truth, the whole truth and nothing but the truth. We will never be perfect, but we shall nonetheless strive towards that goal,” he said on X following the launch.

    Grokipedia’s release had been marked down for the end of September, but was delayed by the US entrepreneur to “purge out the propaganda,” Musk said in a separate X post.

    Musk has been a regular critic of Wikipedia. In 2024, he accused the site of being “controlled by far-left activists” and called for donations to the platform to cease.

    In August, he said “Wikipedia cannot be used as a definitive source for Community Notes, as the editorial control there is extremely left-biased.”

    The content of Grokipedia is generated by artificial intelligence (AI) and the generative AI assistant Grok.

    A Grokipedia article dedicated to Musk states that the Tesla and SpaceX CEO has “influenced broader debates on technological progress, demographic decline, and institutional biases, often via X,”  amid what the page says are “criticisms from legacy media outlets that exhibit systemic left-leaning tilts in coverage.”

    Created in 2001, Wikipedia is a collaborative encyclopedia managed by volunteers, largely funded by donations, and whose pages can be written or edited by internet users.

    It claims a “neutral point of view” in its content.

    (FRANCE 24 with AFP)

  • OpenAI Unveils Search Browser In Challenge to Google

    OpenAI Unveils Search Browser In Challenge to Google

    ChatGPT-maker OpenAI on Tuesday announced an “Atlas” search browser, leveraging its artificial intelligence prowess in a direct challenge to Google Chrome.

    “This is an AI-powered web browser built around ChatGPT,” OpenAI chief Sam Altman said in a streamed presentation.

    OpenAI has ramped up its challenge to Google, which has responded by rapidly building more AI capabilities into search and across its platform.

    Altman and a team of executives demonstrated an “agent” mode that has a chatbot conduct searches on a user’s behalf.

    Altman said that in agent mode, ChatGPT uses the web browser independently, returning with what it finds.

    “It’s got all your stuff and is clicking around,” Altman said.

    “You can watch it or not, you don’t have to, but it’s using the internet for you.”

    Atlas will go live Tuesday on computers powered by Apple’s operating system free of charge, but agent mode will only be available to users of paid Plus or Pro versions of ChatGPT, according to Altman.

    “We want to bring this to Windows and to mobile devices as quickly as we can,” Altman said, without providing a timeline.

    “This is still early days for this project.”

    Some Atlas offerings demonstrated in the stream seemed similar to features already incorporated into Google Chrome and Microsoft Edge internet search browsers.

    – Pressure on Google –

    Tech industry rivals Amazon, Google, Meta, Microsoft and Elon Musk’s xAI have been pouring billions of dollars into artificial intelligence since the blockbuster launch of the first version of ChatGPT in late 2022.

    “OpenAI’s browser puts pressure on Google,” Emarketer technology analyst Jacob Bourne told AFP.

    “This is another step in the AI race as tech companies try to make their AI interfaces the first point of contact for internet users.”

    OpenAI has an opportunity to ride the popularity of ChatGPT to win people over to its browser, according to the analyst.

    However, Bourne noted that Google has a significant infrastructure advantage in terms of providing browser capabilities to billions of users.

    A big question is how well Atlas will perform when under pressure from the kinds of user volume handled by Google, he added.

    The debut of Atlas comes on the heels of Google escaping a breakup of its Chrome browser in a major US competition case, but with the judge imposing remedies whose impact remains uncertain just as AI starts to compete with search engines.

    Judge Amit Mehta, who found a year ago that Google illegally maintained monopolies in online search, did not order the company to sell off its widely-used Chrome browser.

    Instead, he ordered remedies including requirements to share data with other firms so they could develop their own search products, and barring exclusive deals to make Google the only search engine on a device or service.

    Mehta himself noted that the landscape has changed since the US Justice Department and 11 states launched their antitrust case against Google in 2020.

    – Challenges –

    OpenAI, Perplexity, and Microsoft have been ramping up challenges to Google, which dominates the online search market where it earns most of its revenue through targeted advertising.

    OpenAI recently unveiled a new feature for ChatGPT, the leading generative AI model with 800 million weekly users, enabling it to interact with everyday apps like Spotify and Booking.com.

    The new functionality enables ChatGPT to interact with various apps to select music, search for real estate or explore hotel and flight booking sites.

    Meanwhile, Perplexity AI in August announced a new model for sharing search revenue with publishers.

    The company’s media partners will get paid when their work is used by Perplexity’s Comet browser or AI assistant to satisfy queries or requests, according to the San Francisco-based startup.

    Perplexity is one of Silicon Valley’s hottest startups, whose AI-powered search engine is often mentioned as a potential disruptor to Google.

    Google shares were down slightly more than one percent in trading that followed OpenAI announcing Atlas.

    (AFP)

  • Kenya’s John Tingoi Shines Globally, Finishes 3rd in WorldQuant’s 2025 International Quant Championship

    Kenya’s John Tingoi Shines Globally, Finishes 3rd in WorldQuant’s 2025 International Quant Championship

    Nairobi, Kenya – October 6, 2025 — Kenya’s John Tingoi from Chuka University has brought pride to the country after emerging third in the prestigious WorldQuant International Quant Championship (IQC) 2025, a global competition that attracts the world’s top quantitative finance minds.

    Tingoi tied for third place alongside Chia-Chun Chung from Taiwan’s National Yang Ming Chiao Tung University, earning recognition among the top 0.02% of participants worldwide.

    The global challenge, organized by WorldQuant, a leading quantitative asset management firm, drew nearly 80,000 participants from 11,000 universities across 142 countries, who submitted over 263,000 predictive models, known as “alphas.”

    The grand finals were held in Singapore from September 28 to 30, featuring twelve elite teams competing for a share of the $100,000 prize pool.

    The 2025 championship was won by MinKyeom Kim from Ulsan National Institute of Science and Technology (South Korea), with Sumit Kumar from the Indian Institute of Technology Dhanbad (India) coming in second.

    WorldQuant founder and CEO Igor Tulchinsky praised the growing diversity of the competition, noting that this year’s finalists represented four continents.

    “Talent is universal,” he said, adding that the IQC has become a platform that connects education, opportunity, and global collaboration in quantitative finance.

    The International Quant Championship, now in its fifth edition, is powered by WorldQuant’s BRAIN platform, a web-based simulation environment that allows participants to test financial models in real-world conditions.

    The competition aims to democratize access to quantitative finance careers, with top performers often considered for consulting roles, internships, and full-time positions at WorldQuant.

    WorldQuant said that nearly 80 top contestants have so far transitioned into jobs or internships through the IQC and BRAIN programs.

    The 2025 event also introduced Learn2Quant, an educational video series designed to help aspiring quants build essential skills.

    For Kenya, Tingoi’s achievement is a remarkable milestone, placing Chuka University and the country on the global map in the highly competitive field of quantitative research and data-driven finance.

  • iPhone, iPad, iMac: What The ‘i’ in Apple Products Actually Stands For

    iPhone, iPad, iMac: What The ‘i’ in Apple Products Actually Stands For

    With the recent launch of the iPhone 17 series, Apple continues to captivate global markets. However, beyond the sleek designs and advanced features, many users ponder the significance of the lowercase ‘i’ preceding Apple’s product names.

    Initially, during the 1998 iMac debut, Steve Jobs revealed that the ‘i’ stood for ‘internet,’ highlighting Apple’s intent to simplify online connectivity.

    Over time, this meaning has expanded to encompass broader themes, reflecting Apple’s evolving philosophy.

    Five Layers of Meaning

    According to Jobs, the “i” represents more than just internet.

    It conveys individuality, instruction, information, and inspiration, signalling that Apple products are designed to empower users, educate them, and stimulate creativity.

    In this context, the “i” can also be interpreted as the pronoun “I,” emphasising personalisation and user-centric design.

    This layered meaning allowed Apple to extend the “i” branding across a wide range of products, from the iPod in 2001 to the iPhone in 2007 and the iPad in 2010, each product embodying both technological advancement and personal empowerment.

    Evolution in the Age of AI

    With the introduction of AI-powered features in the latest iPhones and iOS updates, the “i” has acquired a new layer of significance: intelligence.

    Apple now uses the branding to signal smart capabilities, from Siri and AI-assisted photography to predictive text and automation, underscoring how the “i” continues to evolve alongside technological progress.

    Beyond its literal meanings, the “i” has become a key branding element for Apple.

    It signifies simplicity, modernity, and accessibility, while reinforcing the identity of Apple products as innovative and user-focused.

    The success of this branding approach has influenced the tech industry, inspiring other companies to adopt minimalist, symbolic naming conventions.

    Apple continues to expand the “i” ecosystem, with rumours of foldable iPads and iPhones in the near future.

    Each new innovation retains the core philosophy symbolised by the “i”: connecting individuals, fostering creativity, and integrating cutting-edge technology.

  • Was It A Hack or Rugpull? Raila Coin Posts On His Social Media Raises Questions

    Was It A Hack or Rugpull? Raila Coin Posts On His Social Media Raises Questions

    A deepfake video and suspicious cryptocurrency promotion on Kenya’s former Prime Minister’s verified X account exposes the growing threat of AI-powered scams in Africa’s crypto space

    Kenya’s crypto community was stunned on September 18 when former Prime Minister Raila Odinga’s verified X account announced the launch of a national cryptocurrency called “Kenya Token.” The post, viewed by millions before its swift deletion, featured a deepfake video of Odinga endorsing what cybersecurity experts quickly identified as an elaborate scam.

    The Viral Announcement That Disappeared

    Raila’s coin post now deleted.
    Raila’s coin post now deleted.

    The deleted post claimed Kenya was “stepping up to lead Africa into the crypto revolution” with a new digital asset built on the Solana blockchain.

    Accompanied by an AI-generated video of Odinga, the message promised the Kenya Token would improve the country’s financial system and support economic growth, directing users to join a Telegram channel for updates.

    However, the informal tone and promotional language were jarringly inconsistent with Odinga’s typical communication style.

    More critically, no government institution corroborated what would have been a major national announcement.

    The Central Bank of Kenya, National Treasury, and other relevant agencies remained silent.

    On Friday, Odinga confirmed the inevitable: “My social media platforms were hacked. The video circulating is fake and misleading. Kindly ignore it.”

    A Growing African Trend

    This incident mirrors a disturbing pattern across the continent.

    In February 2025, Tanzanian billionaire Mohammed Dewji’s X account was similarly compromised to promote a fake “$Tanzania” token.

    Before the account was recovered, scammers had raised nearly $1.48 million through a deepfake endorsement video.

    The Kenya Token project bore hallmarks of hasty assembly. Investigators discovered the project’s website launched on September 17, just one day before the viral post.

    Despite claims of an imminent launch, no contract address existed, making the project impossible to verify or track.

    The Deepfake Challenge

    The sophistication of the AI-generated Odinga video highlights a growing threat.

    While cybersecurity experts identified inconsistencies in voice patterns and lip synchronization, the quality was sufficient to initially fool many observers. This underscores the urgent need for improved digital literacy as deepfake technology becomes more accessible to malicious actors.

    Kenya’s crypto landscape adds another layer of confusion.

    The country already hosts the Kenya Digital Token (KDT), launched in July 2025, though this legitimate project faces its own scrutiny with a single wallet controlling 60% of token supply.

    Virtual Asset Service Providers Bill

    Kenya’s government has been working toward comprehensive cryptocurrency regulation.

    The Virtual Asset Service Providers Bill, scheduled for second reading on September 23, aims to require crypto firms to establish local offices and comply with regulatory oversight.

    The Raila Token incident may accelerate these discussions, demonstrating how easily bad actors can exploit public figures and create market confusion.

    The Central Bank of Kenya maintains a cautious stance toward cryptocurrencies while exploring a Central Bank Digital Currency since 2022.

    The government has also introduced incentives for legitimate crypto adoption, reducing the digital assets trade levy to 1.5% in the 2025 Finance Bill.

    Lessons for Africa’s Crypto Future

    This incident reveals critical vulnerabilities in Africa’s growing crypto ecosystem.

    Even verified accounts of prominent political figures remain susceptible to sophisticated attacks, while AI-generated content becomes increasingly convincing.

    The absence of clear regulatory frameworks creates space for fraudulent projects to operate, particularly in regions where cryptocurrency adoption is rapidly expanding.

    For users, the case emphasizes the importance of verifying announcements through multiple official channels and understanding warning signs of deepfake content.

    The reliance on Telegram for official communications and the lack of technical documentation should have raised immediate red flags.

    For regulators and platforms, the incident demonstrates the need for enhanced security measures, improved deepfake detection capabilities, and faster response mechanisms for fraudulent activities.

    As Kenya and other African nations develop digital currency strategies, protecting citizens from sophisticated scams becomes increasingly critical.

    The continent’s crypto revolution shows immense promise, from Nigeria’s eNaira to Kenya’s potential expansion of M-Pesa into blockchain technology.

    However, the Raila Token incident serves as a sobering reminder that technological advancement must be matched by robust security measures and public education.

    Whether this was purely a security breach or part of a broader rug-pull strategy remains unclear.

    What’s certain is that as Africa embraces digital currencies, the sophistication of threats targeting the ecosystem continues to evolve.

    The incident highlights how quickly public trust can be exploited in the fast-moving world of cryptocurrency, making vigilance and verification more important than ever.

  • Trump Says US Has A Buyer For TikTok

    Trump Says US Has A Buyer For TikTok

    President Donald Trump on Tuesday announced an agreement between the U.S. and China to keep TikTok operating in the United States, with three sources familiar with the matter saying the deal was similar to one discussed earlier this year.

    The agreement requires TikTok’s American assets to be transferred to U.S. owners from China’s ByteDance, potentially resolving a saga that has lingered for nearly a year.

    A deal for the popular social media app, which counts 170 million U.S. users, would represent a breakthrough in months-long talks between the two biggest economies as they seek to defuse a wide-ranging trade war that has unnerved global markets.

    “We have a deal on TikTok … We have a group of very big companies that want to buy it,” Trump said at a White House briefing, without providing further details. The announcement comes a day before a September 17 deadline to sell or shut down the short video app.

    The basics of the new deal, also similar to April, include that ByteDance will keep the single largest ownership stake at 19.9%, just under the law’s 20% threshold, two of the sources said.

    While the broad terms are expected to remain the same, the sources did say they do not know what the final deal would exactly look like, given the potential for last-minute changes.

    U.S. Treasury Secretary Scott Bessent told CNBC on Tuesday the commercial terms of the deal had, in essence, been done since around March with just a few details left to be ironed out.

    “This deal wouldn’t be done without proper safeguards for U.S. national security,” Bessent said. “It seems as though we were also able to meet the Chinese interest.”

    CNBC reported Tuesday that the deal is expected to be closed within the next 30 to 45 days, and that the agreement will include existing investors in TikTok’s China-based parent, ByteDance, and new investors.

    The details are in line with Reuters’ reporting in April that the deal would spin off TikTok’s U.S. operations into a new company based in the U.S. and majority-owned and operated by U.S. investors.
    Advertisement · Scroll to continue

    Any agreement may require approval by the Republican-controlled Congress, which passed a law in 2024 during the Biden Administration that required TikTok’s divestiture due to fears that its U.S. user data could be accessed by the Chinese government, allowing Beijing to spy on Americans or conduct influence operations through the app.

    The Trump administration has declined to enforce the law due to worries it would anger TikTok’s huge user base and disrupt political communications, instead extending the divestiture deadline on three separate occasions.

    Trump, who has credited TikTok with helping him win re-election last year and has 15 million followers on his personal account, was expected to extend the deadline for the fourth time. The White House also launched an official TikTok account last month.

    TARIFFS AND TIKTOK

    A deal for TikTok, which had been in the works in the spring, was put on hold after China indicated it would not approve it following Trump’s announcements of tariffs on Chinese goods.

    Washington has said that TikTok’s ownership by ByteDance makes it beholden to the Chinese government.

    But the company has said U.S. officials have misstated its ties to China, arguing its content recommendation engine and user data are stored in the U.S. on cloud servers operated by Oracle, while content moderation decisions that affect American users are also made in the U.S.

    CNBC reported on Tuesday that Oracle will keep its cloud deal with TikTok. Reuters reported earlier this year that the White House was working on a plan to tap Oracle, along with a group of outside investors, to control the app’s operations.

    As part of the plan, Oracle would have been responsible for addressing national security issues, Reuters had reported.

    Oracle shares pared some gains on Tuesday following the news and were last up 1%.

    A framework agreement was reached by officials from both countries on Monday. A final confirmation on the deal is expected on Friday in a call between Trump and Chinese President Xi Jinping.
    Trump said in March that his administration was in touch with four different groups on TikTok’s sale. Microsoft, Amazon, billionaire Frank McCourt and a consortium led by OnlyFans founder have been among the bidders, according to reports.
    (BBC)
  • Apple Holds Down New iPhone Prices

    Apple Holds Down New iPhone Prices

    CUPERTINO, California, Sept 9 (Reuters) – Apple on Tuesday introduced an upgraded line of new iPhones, including a slimmer iPhone Air, and held prices steady amid U.S. President Donald Trump’s tariffs that have hurt the company’s profit.

    The iPhone Air comes with a high-density battery and a brand new processor, and will be priced starting at $999. Wall Street had been watching to see whether the company would increase the price of iPhones, or seek alternative routes to make up for tariff costs, such as increasing the cost of iPhone versions with more storage.

    The model was the star of the company’s annual product launch event, with CEO Tim Cook saying “we’re taking the biggest leap ever for iPhone.” Apple called it the most durable iPhone yet. It will use the A19 Pro, a processor chip that it said it had tweaked for better energy efficiency and performance to match the device’s thinner battery. Rivals including Samsung already sell phones marketed for their extremely thin profile.
    iPhone Air.
    iPhone Air.
    The iPhone Air’s price is slotted in between the company’s other models, as analysts had predicted. Apple also launched the iPhone 17 and iPhone 17 Pro, the latest upgrade to its flagship smartphone, as well as a new version of its AirPods Pro wireless headphones and a blood pressure monitor in its latest Apple Watch.

    A 256-gigabyte version of the iPhone 17 base model will start at $799, the same as the previous iPhone 16 model with half the storage space. The iPhone 17 Pro will start at $1,099 for a 256-gigabyte model, the same as the previous year’s model with the same storage size, but without the option of a smaller-capacity phone at a lower $999 like the iPhone 16 Pro.

    Apple did not raise the price on watch models or the new AirPods Pro 3 either.

    The prices were an indication that Apple was ready to swallow the cost of tariffs to ward off competition from Samsung, Alphabet’s Google and domestic rivals in China, even as it has predicted that the levies would cost it more than $1 billion in the current fiscal quarter.

    Apple shares were down 1.6% after the company announced pricing of the smartphones.

    “They’re leveraging their scale to try to keep prices where they are,” said Tom Mainelli, head of IDC’s Device & Consumer Research Group. “I think Apple, like most tech vendors, are acutely aware, particularly in the U.S., that tariffs are going to impact consumers’ ability to spend in the second half of this year. So … they’re going to hold a line on prices and try to make it possible for people to upgrade between now and the end of the year at the same price as last year.”

    The iPhone Air will go head-to-head against Samsung Electronics’ Galaxy S25 Edge, and analysts told Reuters it could be a stepping stone toward competing with Samsung’s folding phones, which are in their seventh generation. A foldable phone is important for Apple to appeal to customers in China, where consumers like foldables and the company has been losing market share.

    “This new device will bring a sense of newness to the iPhone, which has remained the same for too long,” said PP Foresight analyst Paolo Pescatore. He said the “new and much-improved iPhone line-up looks impressive, which puts (Apple) in a strong position to cater for different segments.”

    BLOOD PRESSURE MONITOR

    The event was light on commentary on how Apple aimed to close the gap with the likes of Google, which has used its latest flagship phones to showcase the capabilities of its Gemini AI models. The company has leaned on a partnership with ChatGPT creator OpenAI to power many AI features on its devices.

    Apple sidestepped “the heart of the AI arms race while positioning itself as a longtime innovator on the AI hardware front, with silicon and device-level integration,” said eMarketer analyst Gadjo Sevilla.

    Apple said the iPhone Air will also feature a new “N1” chip to handle Wi-Fi communications and new “C1X” modem for cellular data. Chips for those functions in Apple’s premium devices were long supplied by Broadcom and Qualcomm, whose shares were down 2.3% and 1.1%, respectively, late on Tuesday afternoon.

    “This is MacBook Pro levels of compute, in an iPhone,” Tim Millet, one of Apple’s chip executives, said during the presentation at the company’s Cupertino, California, headquarters. The iPhone Air will have two cameras and eliminate the physical SIM card slot, freeing up more room for battery capacity.

    Apple said the base model iPhone 17 will have a brighter, more scratch-resistant screen. It will also have a better front-facing camera with a differently shaped sensor to make horizontal selfies look better.

    The new AirPods Pro 3 will feature live translation of languages. Apple also said that if both people in a conversation are wearing the new AirPods Pro 3, the earbuds will translate conversations in near real time.

    The blood pressure monitor feature is pending regulatory approval, Apple said. The watch will not detect every case of high blood pressure but the company said it expects the feature to notify 1 million people and will make it available in 150 countries.

  • Apple Unveils iPhone 17 Air Amid AI Race and Tariff Pressures

    Apple Unveils iPhone 17 Air Amid AI Race and Tariff Pressures

    Apple unveiled its iPhone 17 lineup on Tuesday, featuring its thinnest smartphone ever, as the tech giant works to prove it can keep pace in the generative AI race.

    The Silicon Valley powerhouse held its annual iPhone release event amid mounting pressures: the White House is pushing the company to reduce its dependence on Chinese manufacturing, while investors question whether Apple is truly ready for the AI age.

    Adding to these challenges, the company faces headwinds from President Donald Trump’s high tariff policies. Apple shares have dropped more than three per cent since the Republican took office in January.

    Against this backdrop, Apple is betting on a product that it hopes will spark a super-cycle of iPhone purchases and reverse the trend of customers holding onto their devices longer before upgrading.

    To reinvigorate its brand, Apple introduced the iPhone 17 Air, which CEO Tim Cook called “a total game changer.”

    At just 5.6mm thick, the $999 device features Apple’s new A19 Pro processor — its most powerful iPhone chip to date — and promises all-day battery life with up to 40 hours of video playback.

    The Air joins Apple’s standard lineup, including the premium iPhone Pro 17, the company’s most expensive and highest-performing model.

    While all new devices incorporate generative AI technology, Apple made no major announcements about expanding its AI capabilities beyond updates to existing features in its “Apple Intelligence” suite.

    The company’s AI push has struggled to gain traction since it launched “Apple Intelligence” late last year. Users have been particularly disappointed with improvements to Siri, which remains surprisingly basic despite years of promises.

    Apple reportedly plans to integrate AI into online search next year and overhaul Siri, though the company hasn’t confirmed these reports.

    The tech giant is also said to be partnering with Google to tap into its search and AI expertise.

    Pivot to Air

    Industry analysts see the iPhone Air as a strategic pivot, with Apple positioning ultra-thin design — rather than larger screens — as its new premium selling point.

    The super-slim profile could also pave the way for Apple’s long-rumoured foldable iPhone, expected as early as next year. Samsung and China’s Huawei already offer foldable smartphones.

    However, engineering such thin devices presents challenges: higher production costs and reduced battery space, though Apple claims 24-hour battery life for the iPhone 17 Air when fully charged.

    Despite tariffs that increase production costs, Apple kept iPhone prices unchanged from last year’s equivalent models — a move that may squeeze profit margins.

    Cook revealed in July that Trump’s tariffs cost Apple $800 million last quarter, with an estimated $1.1 billion impact expected this quarter.

    Apple shares fell 1.35 per cent shortly after pricing details were announced, reflecting investor concerns about the company’s ability to maintain its profit margins.

    Apple also introduced the AirPods Pro 3, featuring enhanced noise cancellation and real-time translation capabilities, along with the Apple Watch Series 11, which includes 5G connectivity, extended battery life, and heart health monitoring features pending regulatory approval.

    (AFP)

  • Apple iPhone Event May Lack Sparkle, But Rumored iPhone Air Likely to Spur Upgrades

    Apple iPhone Event May Lack Sparkle, But Rumored iPhone Air Likely to Spur Upgrades

    SAN FRANCISCO, Sept 8 (Reuters) – When Apple  debuts new iPhones on Tuesday, analysts say its biggest challenge will be to ride out another ho-hum launch as rivals have skated past it in embedding artificial intelligence into their products and services.

    The biggest draw this year could be a rumored “iPhone Air”, a phone slimmer than what Apple has sold before and taking its name from the company’s slender laptop MacBook Air.

    Apple would need to iron out how to pack batteries and cameras into a thinner device, analysts said, and seek to price it between the base iPhone 17 models and more expensive Pro models to attract a large number of customers.

    Dipanjan Chatterjee, vice president and principal analyst at Forrester, said a slimmer iPhone could spur upgrades. “It’s been a while since we have seen any meaningful update to the form factor of the device beyond tepid incremental changes, and the novelty of the Air will likely induce many 14, 15 and even 16 iPhone users to migrate up,” Chatterjee said.

    The slimmer phone could also be a stepping stone toward an iPhone that folds out flat like a book and would act as a platform for an upgraded Siri, neither of which are likely to arrive until next year, analysts said.

    Samsung Electronics is on its seventh generation of folding phones and Alphabet’s is Google on its third, yet Chatterjee estimates they are less than 2% of all phone sales and will not grow beyond 5% “anytime soon.”

    A foldable phone is important for Apple to appeal to its customers in China, where consumers like foldables and the company has been losing market share, analysts said.

    Historically, Cupertino, California-based Apple has received nearly a quarter of its total sales from the mid-price iPhone segment, said Gene Munster, managing partner at Deepwater Asset Management, which holds Apple shares. He expects Apple to find a way to raise prices across its iPhone lineup without any allusion to U.S. President Donald Trump’s tariffs, perhaps by raising prices for larger storage capacity.

    “They learned to play nice with Washington – a straight up price increase might not come off well,” Munster said. “But I think they have rising costs, and they have been loyal to growing margins, and to do that, you have to find some method.”

    AI STRUGGLES

    Initially, Apple slated improvements to Siri for last spring, but delayed them due to engineering setbacks. Instead, it leaned on a partnership with ChatGPT creator OpenAI to power many AI features on its devices. In contrast, Google’s latest flagship phones were largely designed to showcase the capabilities of its Gemini AI models.

    Apple is also in early talks to use Google’s Gemini AI to revamp Siri, according to a recent media report.

    Ben Bajarin, CEO of technology consultancy Creative Strategies, said Apple may highlight improved AI processing capabilities of its next generation of Apple Silicon chips that could, in the future, power an “agentic” Siri that takes care of tasks in the background for the owners of the 2.35 billion Apple devices in use around the world, without draining the device’s battery.

    “It could be foreshadowing of a broader kind of agentic integration with their operating system, because the operating system is going to be the thing that hits the (chip’s AI processing capabilities) the most,” Bajarin said.

    Bob O’Donnell, president at TECHnalysis Research, said that while AI adoption remains in early stages, Apple’s time to catch up without suffering a long-term setback is now likely measured in months rather than years.

    “They have a huge share in the U.S., and most people are perfectly content,” he said. “But by this time next year, if Siri still sucks, and if they don’t get the foldable out, I don’t know” whether that contentment will continue, O’Donnell said.

  • OpenAI Launches Its GPT-5 Artificial Intelligence Model

    OpenAI Launches Its GPT-5 Artificial Intelligence Model

    OpenAI launched on Thursday its GPT-5 artificial intelligence model, the highly anticipated latest installment of a technology that has helped transform global business and culture.

    OpenAI’s GPT models are the AI technology that powers the popular ChatGPT chatbot, and GPT-5 will be available to all 700 million ChatGPT users, OpenAI said.

    The big question is whether the company that kicked off the generative AI frenzy will be capable of continuing to drive significant technological advancements that attract enterprise-level users to justify the enormous sums of money it is investing to fuel these developments.

    The release comes at a critical time for the AI industry. The world’s biggest AI developers – Alphabet, Meta, Amazon and Microsoft, which backs OpenAI – have dramatically increased capital expenditures to pay for AI data centers, nourishing investor hopes for great returns. These four companies expect to spend nearly $400 billion this fiscal year in total.

    OpenAI is now in early discussions to allow employees to cash out at a $500 billion valuation, a huge step-up from its current $300 billion valuation. Top AI researchers now command $100 million signing bonuses.

    “So far, business spending on AI has been pretty weak, while consumer spending on AI has been fairly robust because people love to chat with ChatGPT,” said economics writer Noah Smith. “But the consumer spending on AI just isn’t going to be nearly enough to justify all the money that is being spent on AI data centers.”

    OpenAI is emphasizing GPT-5’s enterprise prowess. In addition to software development, the company said GPT-5 excels in writing, health-related queries, and finance.

    “GPT-5 is really the first time that I think one of our mainline models has felt like you can ask a legitimate expert, a PhD-level expert, anything,” OpenAI CEO Sam Altman said at a press briefing.

    “One of the coolest things it can do is write you good instantaneous software. This idea of software on demand is going to be one of the defining features of the GPT-5 era.”

    In demos on Thursday, OpenAI showed how GPT-5 could be used to create entire working pieces of software based on written text prompts, commonly known as “vibe coding.”

    One key measure of success is whether the step up from GPT-4 to GPT-5 is on par with the research lab’s previous improvements. Two early reviewers told Reuters that while the new model impressed them with its ability to code and solve science and math problems, they believe the leap from the GPT-4 to GPT-5 was not as large as OpenAI’s prior improvements.

    Even if the improvements are large, GPT-5 is not advanced enough to wholesale replace humans. Altman said that GPT-5 still lacks the ability to learn on its own, a key component to enabling AI to match human abilities.

    On his popular AI podcast, Dwarkesh Patel compared current AI to teaching a child to play a saxophone by reading notes from the last student.

    “A student takes one attempt,” he said. “The moment they make a mistake, you send them away and write detailed instructions about what went wrong. The next student reads your notes and tries to play Charlie Parker cold. When they fail, you refine the instructions for the next student. This just wouldn’t work.”

    MORE THINKING

    Nearly three years ago, ChatGPT introduced the world to generative AI, dazzling users with its ability to write humanlike prose and poetry, quickly becoming one of the fastest growing apps ever.

    In March 2023, OpenAI followed up ChatGPT with the release of GPT-4, a large language model that made huge leaps forward in intelligence. While GPT-3.5, an earlier version, received a bar exam score in the bottom 10%, GPT-4 passed the simulated bar exam in the top 10%.

    GPT-4’s leap was based on more compute power and data, and the company was hoping that “scaling up” in a similar way would consistently lead to improved AI models.

    But OpenAI ran into issues scaling up. One problem was the data wall the company ran into, and OpenAI’s former chief scientist Ilya Sutskever said last year that while processing power was growing, the amount of data was not.

    He was referring to the fact that large language models are trained on massive datasets that scrape the entire internet, and AI labs have no other options for large troves of human-generated textual data.

    Apart from the lack of data, another problem was that ‘training runs’ for large models are more likely to have hardware-induced failures given how complicated the system is, and researchers may not know the eventual performance of the models until the end of the run, which can take months.

    At the same time, OpenAI discovered another route to smarter AI, called “test-time compute,” a way to have the AI model spend more time compute power “thinking” about each question, allowing it to solve challenging tasks such as math or complex operations that demand advanced reasoning and decision-making.

    GPT-5 acts as a router, meaning if a user asks GPT-5 a particularly hard problem, it will use test-time compute to answer the question.

    This is the first time the general public will have access to OpenAI’s test-time compute technology, something that Altman said is important to the company’s mission to build AI that benefits all of humanity.

    Altman believes the current investment in AI is still inadequate.

    “We need to build a lot more infrastructure globally to have AI locally available in all these markets,” Altman said.

    (Reuters)

  • Microsoft Servers Hacked By Chinese Groups, Firm Says

    Microsoft Servers Hacked By Chinese Groups, Firm Says

    Chinese “threat actors” have hacked Microsoft’s SharePoint document software servers and targeted the data of the businesses using it, the firm has said.

    China state-backed Linen Typhoon and Violet Typhoon as well as China-based Storm-2603 were said to have “exploited vulnerabilities” in on-premises SharePoint servers, the kind used by firms, but not in its cloud-based service.

    The US tech giant has released security updates in response and has advised all on-premises SharePoint server customers to install them.

    “Investigations into other actors also using these exploits are still ongoing,” Microsoft said in a statement.

    The firm said it had “high confidence” the hackers would continue to target systems which have not installed its security updates.

    It added that it would update its website blog with more information as its investigation continues.

    Microsoft said it had observed attacks in which hackers had sent a request to a SharePoint server “enabling the theft of the key material by threat actors”.

    Charles Carmakal, chief technology officer at Mandiant Consulting firm, a division of Google Cloud, told the BBC it was “aware of several victims in several different sectors across a number of global geographies”.

    Carmakal said it appeared that governments and businesses that use SharePoint on their sites were the primary target.

    A number of adversaries who stole material encoded by cryptography were then able to regain ongoing access to the victims’ SharePoint data, he said.

    “This was exploited in a very broad way, very opportunistically before a patch was made available. That’s why this is significant,” Carmakal said.

    Carmakal said the “China-nexus actor” was deploying techniques similar to previous campaigns associated with Beijing.

    Microsoft said Linen Typhoon had “focused on stealing intellectual property, primarily targeting organizations related to government, defence, strategic planning, and human rights” for 13 years.

    It added that Violet Typhoon had been “dedicated to espionage”, primarily targeting former government and military staff, non-governmental organizations, think tanks, higher education, the media, the financial sector and the health sector in the US, Europe, and East Asia.

    Meanwhile, Storm-2603 was “assessed with medium confidence to be a China-based threat actor”.

    (BBC)

  • Musk’s Grok Signs $200m Deal With Pentagon

    Musk’s Grok Signs $200m Deal With Pentagon

    The Pentagon has signed a multi-million dollar deal to begin using Elon Musk’s artificial intelligence chatbot, Grok, as part of a wider rollout of AI tools for government use, the Department of Defence confirmed.

    Announced on Monday by Musk’s company xAI, the $200m (£149m) contract is part of its “Grok for Government” programme, and aligns with the Trump administration’s push for more aggressive adoption of artificial intelligence.

    It comes just days after Grok sparked backlash for spouting antisemitic posts, including praise for Adolf Hitler on X, the social media platform owned by Musk.

    Musk said the bot was “too compliant” and “too eager to please”. He said the issue was being addressed.

    Musk’s xAI says the new deal will give US government departments access to Grok 4, the latest version of the chatbot, and offer custom tools for national security use.

    The company also plans to provide technical support for classified environments.

    The Pentagon also announced awarding similar contracts to Anthropic, Google and OpenAI – each with a $200m ceiling.

    “The adoption of AI is transforming the Department’s ability to support our warfighters and maintain strategic advantage over our adversaries,” said the administration’s Chief Digital and AI Officer Doug Matty.

    Musk’s expanding government partnerships come amid a deteriorating relationship with President Donald Trump.

    The Tesla and SpaceX boss had spent a quarter of a billion dollars on Trump’s re-election effort in 2024, and actively campaigned for him.

    He was later appointed to run the Department of Government Efficiency (Doge) – a federal cost-cutting initiative tasked with reducing the size of the US government.

    But in recent months, Musk began openly criticising what Trump had dubbed the “Big Beautiful Bill”, a sprawling spending and tax cuts legislation that the Tesla boss said was too costly for Americans.

    Musk resigned from his post at Doge in May, though the department has not been officially disbanded.

    Since then, Trump had suggested Doge could be deployed to harm Musk’s companies.

    Trump also suggested he might deport Musk, who is an American citizen and was born in South Africa. He also holds Canadian citizenship.

    While at the helm of Doge, the White House was criticised for allowing Musk to have unfettered access to troves of government data on American citizens.

    Despite the fall-out, Musk’s xAI has continued to expand its government work. Its newly-announced contract may also create an avenue for that data collection to continue.

    Grok was introduced in late 2023 as a more unfiltered alternative to other AI chatbots like ChatGPT. It is already integrated into Musk’s social media platform X, formerly known as Twitter.

    (BBC)

  • Details: How MPs Plot to Regulate Artificial Intelligence Usage in Kenya

    Details: How MPs Plot to Regulate Artificial Intelligence Usage in Kenya

    Kenyan lawmakers are intensifying their push for comprehensive regulation of artificial intelligence (AI) usage in the country, citing growing concerns over disinformation, privacy breaches, and potential national security threats posed by unregulated AI systems.

    At the heart of this legislative initiative is a motion by Aldai MP Marianne Kitany, who has called for the formulation of a robust regulatory framework to govern AI implementation across Kenya.

    The motion, which has gained significant traction among parliamentarians, specifically requests the government, through the Ministry of Information, Communication and the Digital Economy, to develop comprehensive regulatory guidelines and ethical standards for AI use.

    “Government, through the Ministry of Information, Communication and the Digital Economy, should formulate a regulatory framework and ethical guidelines for implementation of Artificial Intelligence in the country to control its potential misuse,” reads Kitany’s motion, which was presented to the National Assembly.

    The lawmakers’ push for regulation stems from mounting concerns about AI’s potential for misuse in a country where digital literacy is still developing.

    MPs have pointed to the proliferation of fake news and disinformation campaigns, many of which are now AI-powered, as primary reasons for urgent regulatory intervention.

    Nambale MP Geoffrey Mulanya highlighted the immediate dangers facing Kenyan society.

    “We have had cases of fake news appearing in our social media because we have young intelligent people coming up with screaming fake newspaper headlines, which causes conflict in our society,” Mulanya told the National Assembly.

    The concerns extend beyond disinformation to encompass a broader range of AI-related risks that MPs believe could destabilize the country if left unchecked.

    These include algorithmic discrimination, privacy invasion, financial market manipulation, job displacement, and the potential development of autonomous weapons systems.

    While acknowledging AI’s transformative potential across various sectors, lawmakers have emphasized the technology’s double-edged nature.

    The MPs recognize that AI has already brought significant benefits to Kenya’s healthcare, manufacturing, and robotics sectors, improving efficiency and driving innovation.

    Luanda MP Dick Maungu noted the inevitability of AI adoption in Kenya’s increasingly connected society.

    “We live in a society which is a global village. If we don’t have a regulatory framework, people are set to suffer,” Maungu warned, emphasizing the urgent need for proactive legislation.

    Bondo MP Gideon Ochanda drew parallels with previous technological adoptions, citing how Kenyans initially resisted tea-picking machines but eventually embraced them once their benefits became apparent.

    “Artificial Intelligence is a must, and if we don’t regulate it, it’s going to run ahead of us, ahead of the government, and people are going to continue using it, and they are already doing that,” Ochanda stated.

    The parliamentary motion outlines a multi-faceted approach to AI regulation that goes beyond mere restriction to include education and awareness components.

    MPs want the government to develop and execute a public awareness program on AI to increase understanding of the technology, foster transparency, and promote responsible usage.

    MP Erick Muchangi emphasized the urgency of regulatory action, pointing to AI’s growing presence in critical sectors.

    “The government must move fast and regulate AI, because it is nowadays being used in healthcare, in the education sector, so how can we not regulate it?” he questioned.

    West Mugirango MP Stephen Mogaka has called for AI to be integrated into the curriculum of technical training institutions, signaling lawmakers’ recognition that regulation must be coupled with education and skill development.

    A key concern among MPs is ensuring that AI regulation doesn’t stifle innovation while protecting employment opportunities for Kenya’s youth.

    The lawmakers have cautioned against embracing AI “100 percent,” warning that unrestricted adoption could lead to significant job losses, particularly among young people who form a substantial portion of Kenya’s workforce.

    This balanced approach reflects the MPs’ understanding that Kenya must navigate between harnessing AI’s benefits and protecting its citizens from potential negative consequences.

    The parliamentary discussions have also highlighted national security implications of unregulated AI use.

    MPs have emphasized the need for comprehensive security assessments before allowing AI systems access to critical national data, recognizing that improper AI implementation could compromise Kenya’s digital sovereignty and security infrastructure.

    The push for regulation comes as Kenya seeks to improve its position in global AI adoption rankings.

    According to the 2022 Government Artificial Intelligence Readiness Index report, Kenya ranked fifth in Africa and 90th globally in AI adoption readiness.

    The Oxford Insights Survey 2022 placed Kenya’s AI readiness at 40.3 percent, indicating significant room for improvement.

    The Ministry of Information, Communication and the Digital Economy has been working on Kenya’s National AI Strategy 2025-2030, which aims to position the country as Africa’s leading AI hub.

    The Kenyan Ministry of Information, Communications, and the Digital Economy has released a draft National AI Strategy for public validation.

    The State Department for Parliamentary Affairs has reaffirmed its commitment to coordinating the legislative agenda that will anchor Kenya’s National AI Strategy 2025–2030 in law, ensuring a strong regulatory framework for artificial intelligence adoption.

    Currently, there are currently no specific laws or regulations in Kenya that directly regulate AI, making the MPs’ push for comprehensive legislation particularly timely.

    The Kenya AI Strategy 2025–2030 positions itself as one of the most structured and forward-looking national frameworks in sub-Saharan Africa to date, suggesting that Kenya’s regulatory framework could serve as a model for other African nations grappling with similar AI governance challenges.

    The parliamentary motion represents a critical step in Kenya’s journey toward responsible AI adoption, balancing the need for innovation with the imperative to protect citizens from potential AI-related harms.

    As the debate continues in the National Assembly, the outcome could significantly influence how Kenya navigates the complex landscape of AI governance in the coming years.

    The MPs’ initiative reflects a growing global trend toward AI regulation, with Kenya positioning itself to be among the first African nations to establish comprehensive AI governance frameworks. The success of this legislative effort could determine whether Kenya achieves its ambition of becoming a regional AI leader while maintaining the safety and security of its digital ecosystem.

  • WhatsApp To Start Showing Ads

    WhatsApp To Start Showing Ads

    It took 11 years since Facebook acquired it for $19 billion, but Meta is finally bringing ads to WhatsApp, marking a major change for an app whose founders shunned advertising.

    Meta announced Monday that businesses will now be able to run so-called status ads on WhatsApp that prompt users to interact with the advertisers via the app’s messaging features. The ads will only be shown to users within WhatsApp’s “Updates” tab to separate the promotions from people’s personal conversations. Additionally, Meta will begin monetizing WhatsApp’s Channels feature through search ads and subscriptions.

    The debut of ads on the messaging app represents a significant step in Meta CEO Mark Zuckerberg’s plans to make WhatsApp “the next chapter” in his company’s history, as he told CNBC’s Jim Cramer in 2022.

    The move to monetize WhatsApp also comes amid Meta’s high-profile antitrust case with the Federal Trade Commission over the company’s blockbuster acquisitions of the messaging app and Instagram.

    Already, Meta allows advertisers to run so-called click-to-message ads on Facebook and Instagram that steer users to WhatsApp where they can directly engage with businesses.

    Messaging between brands and consumers “should be the next pillar of our business,” Zuckerberg told analysts in April, adding that WhatsApp now has over 3 billion monthly users, including “more than 100 million people in the U.S. and growing quickly there.”

    Now, companies can run those kinds of ads within WhatsApp itself.

    The new status ads appear in a user’s Updates tab within that tab’s “Status” feature that can be used to share pictures, videos and text that vanish after 24 hours, akin to Instagram Stories.

    Since Meta bought WhatsApp in 2014, the popular messaging app has continued to grow worldwide. But unlike Facebook, Instagram and most recently Threads, WhatsApp has never allowed advertising.

    WhatsApp’s co-founders, Jan Koum and Brian Acton, were public in their scorn for the advertising industry, and the duo left Facebook after reportedly clashing with executives who were eager to inject the app with advertising and other practices they shunned.

    The social media company does not reveal WhatsApp’s specific sales, but analysts have previously estimated the app’s revenue to be between $500 million and $1 billion from charging businesses for tools and services so they can message customers on the app.

    Meta will “use very basic information” to recommend which ads to show WhatsApp users, Nikila Srinivasan, Meta’s head of product for business messaging, said Friday.

    This includes a person’s country, city, device, language and data like who they follow or how they interact with ads.

    The company debuted WhatsApp’s Updates tab in June 2023 along with an accompanying Channels feature that allows people and organizations to send broadcast messages and updates to their followers as opposed to personal conversations.

    Meta will also monetize the Channels feature, the company said Monday.

    Organizations and people who are Channel administrators will now be able to spend money to boost the visibility of their respective Channels when a person searches for them via a directory, similar to ads on Apple’s and Google’s app stores.

    Additionally, channel administrators will be able to charge users monthly subscription fees to access exclusive updates and content, Meta said Monday.

    The company will not immediately make money from those monthly subscription fees, but it plans to eventually take a 10% cut of those subscriptions, a spokesperson said.

    Meta hopes that by limiting its new ads to WhatsApp’s Updates tab it will disrupt users as little as possible, Srinivasan said.

    Users’ status updates as well as personal messages and calls on WhatsApp will remain encrypted, she said.

    “We really believe that the Updates tab is the right place for these new features,” Srinivasan said.

  • OpenAI Wins Ksh.25.9 Billion US Defense Contract

    OpenAI Wins Ksh.25.9 Billion US Defense Contract

    ChatGPT maker OpenAI was awarded a $200 million (approximately Ksh.25.9 billion) contract to provide the U.S. Defense Department with artificial intelligence tools, the Pentagon said in a statement on Monday.

    “Under this award, the performer will develop prototype frontier AI capabilities to address critical national security challenges in both war fighting and enterprise domains,” the Pentagon said.

    OpenAI said last week that its annualized revenue run rate surged to $10 billion as of June, positioning the company to hit its full-year target amid booming AI adoption.

    OpenAI said in March it would raise up to $40 billion in a new funding round led by SoftBank Group (9984.T), at a $300 billion valuation. OpenAI had 500 million weekly active users as of the end of March.

    The White House’s Office of Management and Budget released new guidance in April directing federal agencies to ensure that the government and “the public benefit from a competitive American AI marketplace.”

    The guidance had exempted national security and defense systems.

    (Reuters)

  • New Bill To Allow Govt To See All Websites You’ve Visited

    New Bill To Allow Govt To See All Websites You’ve Visited

    Controversial legislation would require internet providers to track and report all user activity to government authorities

    A new bill making its way through Parliament could give the Kenyan government unprecedented access to citizens’ internet browsing history, sparking concerns about digital privacy and surveillance overreach.

    The Kenya Information and Communications (Amendment) Bill 2025, introduced by Aldai MP Marianne Jebet Kitany, would require all Internet Service Providers to assign unique tracking numbers to customers and submit detailed annual reports of their online activities to the Communications Authority of Kenya.

    Government Database of Internet Activity

    Under the proposed legislation, ISPs would be mandated to monitor and record which websites their customers visit, creating what critics describe as a comprehensive government database of Kenyans’ digital lives.

    The bill specifically requires service providers to “monitor customer usage” and “convert customer usage into readable details.”

    The tracking system would assign each internet user a unique identification number, similar to a utility meter, allowing authorities to link browsing history directly to individual citizens.

    This data would then be compiled annually and handed over to government regulators.

    “This creates an unprecedented level of digital surveillance,” said a digital rights advocate who requested anonymity.

    “The government would essentially have a record of every website visited by every Kenyan with internet access.”

    Expert Warns of Hidden Surveillance Architecture

    Aldai MP Marianne Jebet Kitany is the sponsor of The Kenya Information and Communications (Amendment) Bill 2025.
    Aldai MP Marianne Jebet Kitany is the sponsor of The Kenya Information and Communications (Amendment) Bill 2025.

    Dr. Sarah Mwangi, a cybersecurity expert and digital rights researcher at the University of Nairobi, warns that the bill’s true implications extend far beyond simple internet billing, as the government claims.

    “The requirement for unique meter numbers for every internet user isn’t about billing—it’s about creating a tracking infrastructure,” Dr. Mwangi explained.

    “When you combine individual identification with the mandate to ‘convert internet usage into readable details,’ you’re looking at comprehensive digital surveillance.”

    She raises critical concerns about the bill’s scope: “What does ‘readable details’ actually mean? Does this include which websites you visit, your private messages, your voice calls? The language is deliberately vague, which is deeply troubling from a privacy perspective.”

    Dr. Mwangi particularly warns about the potential for future abuse: “Once you build this kind of digital tracking architecture, it becomes very easy to expand it. We could be looking at the foundation for a social credit system where citizens are scored and potentially punished based on their online behavior.”

    Threat to Journalism and Civil Society

    The cybersecurity expert highlights specific risks for vulnerable groups: “Journalists, activists, and whistleblowers rely on digital anonymity to operate safely. This bill could effectively end anonymous online participation in Kenya, creating a chilling effect on free speech and investigative reporting.”

    She points to concerning questions about data sharing: “The bill doesn’t clearly specify whether this internet usage data will be shared with national security agencies. If it is, we’re talking about mass surveillance of the entire population, not targeted security measures.”

    Social Media Identity Verification Required

    The bill also targets social media platforms, requiring companies like Facebook, WhatsApp, and Instagram to verify users’ ages through national identification documents.

    While presented as a child protection measure, critics argue this would eliminate anonymous online participation and make it easier for authorities to identify social media users.

    Major international platforms would need to implement new verification systems specifically for Kenyan users, raising questions about compliance and enforcement across global technology companies.

    Constitutional Concerns Raised

    Legal experts are questioning whether the proposed surveillance measures violate constitutional protections for privacy and freedom of expression.

    Kenya’s 2010 Constitution guarantees citizens’ right to privacy, including freedom from unreasonable searches of their property and possessions.

    “There are serious constitutional issues with requiring blanket surveillance of all internet users,” explained a constitutional lawyer familiar with the bill.

    “Such broad monitoring powers would typically require judicial oversight and specific justification.”

    The legislation comes amid growing concerns about digital freedom in Kenya. Freedom House’s 2024 report documented increasing online censorship, including 64 content removal requests submitted to Google by Kenyan authorities in 2023.

    Control Over Digital Infrastructure

    Dr. Mwangi raises additional concerns about infrastructure control: “A key question is who will control this metering infrastructure—the government or private telecoms? If it’s state-controlled, the government essentially controls what citizens can see, say, and stream online.”

    She warns about the lack of safeguards against political targeting: “What protections exist against profiling, targeting, or discrimination? What happens when political opponents can be tracked and potentially harassed based on their internet behavior? These are fundamental questions the bill doesn’t address.”

    Technical and Economic Implications

    Internet service providers would face significant new compliance costs under the proposed system. Companies would need to invest in tracking infrastructure, data storage systems, and reporting mechanisms to meet the government’s requirements.

    These costs could potentially be passed on to consumers through higher internet prices, potentially limiting digital access for lower-income Kenyans.

    Technology experts also question the technical feasibility of comprehensive internet monitoring, particularly with the increasing use of encrypted connections and virtual private networks that can mask user activity.

    If passed, Kenya’s legislation could influence similar efforts across East Africa, where governments are grappling with balancing security concerns against digital rights.

    The bill represents one of the region’s most comprehensive attempts at internet surveillance legislation.

    Other countries have implemented various forms of online monitoring, but few have required such extensive tracking of individual browsing habits by private internet companies.

    The bill will undergo committee review and public participation processes before potential passage. Digital rights organizations are preparing submissions opposing the surveillance provisions, while government supporters argue the measures are necessary for national security and child protection.

    Citizens concerned about the legislation can participate in public hearings and submit comments through official parliamentary channels during the review process.

    The outcome will significantly impact how Kenyans access and use the internet, potentially affecting everything from e-commerce to social media participation and online journalism in the country.

    Dr. Mwangi concludes: “Citizens need to pay close attention to this bill. It may be presented as simple internet metering, but the surveillance architecture it creates could fundamentally change the relationship between Kenyans and their government in the digital space.”​​​​​​​​​​​​​​​​