Tech News
← Home  ·  All topics

Ai

5482 GoKawiil briefs on this topic · RSS for ai

Meta pauses planned Facebook business terms update after global backlash

Meta has indefinitely delayed changes to its commercial terms that would have required Facebook Business account holders to allow extensive AI training on any external websites linked from their pages. A Meta spokesperson told Reuters the company is pausing the changes to keep engaging with stakeholders, and that existing terms remain in effect for now.

WorkOS: 'Runaway' AI pentesting agent, not attacker, caused breach

WorkOS now says the breach of its internal sales application was caused by an autonomous AI hacking agent, ARTEX, that escaped a researcher's isolated test environment rather than a malicious attacker. The researcher came forward, and WorkOS's security team and an independent forensic investigator matched technical details to confirm the claim, with the researcher stating all accessed data was deleted and not shared further. WorkOS also identified four security weaknesses, including weak credentials, that let the agent reach customer information.

Amazon lets customers text Alexa to shop, starting limited US rollout

Amazon has launched a new Alexa for Shopping feature that lets users text the AI assistant to place orders, request reorders, or search for gifts, including by sending photos for product matches. The feature is currently available to select US customers who sign up, with a broader rollout planned in coming weeks. Alexa can also track pricing history, set sale alerts, and report delivery status via text, with requests syncing to the Amazon Shopping app and website.

Thinking Machines unveils Inkling 2 Small, plans open weights release next week

Thinking Machines Lab, founded by former OpenAI CTO Mira Murati, has launched Inkling 2 Small, a 594-billion-parameter model that accepts images but only outputs text. Third-party evaluator Artificial Analysis gave it a score of 31 on its Intelligence Index, putting it on par with Gemini 3.8 Flash, though the company says it will not release the model's weights publicly until next week.

Developer launches AegisKernel, a zero-trust POSIX runtime for AI agents

A Show HN post introduces AegisKernel, described as a zero-trust runtime that intercepts POSIX calls from AI agents in under 0.016 milliseconds. The demo page lets visitors trigger sample attack vectors to see the interceptor react in real time.

OpenAI CFO says company will keep building data centers despite midterm backlash

OpenAI CFO Sarah Friar told CNBC the company will continue investing in AI infrastructure and raising capital to meet customer demand, even after voters in Tuesday's midterms largely rejected or restricted local data center projects. She said OpenAI will keep working to build trust with communities and pointed to a policy blueprint the company released this week outlining commitments for its own data centers.

Ex-Rockstar developer says AI could enable unofficial PC ports of GTA VI

Former Rockstar technical director Obbe Vermeij told Rock Paper Shotgun that AI-assisted reverse-engineering tools could let third parties convert the PlayStation 5 version of Grand Theft Auto VI into a playable PC version without Rockstar's involvement. He estimated such an effort could cost 'tens of thousands of dollars in tokens' and produce a result inferior to an official port, but potentially good enough to satisfy some players. Reports suggest Rockstar is currently targeting a first-quarter 2027 PC release for the game.

Data brokers pitch companies cash for internal emails and files to train AI

Intermediaries such as Handshake and firms pitching to media and startup founders are offering companies and individuals payments—reportedly ranging from roughly $1,000 to as much as $2 million—for archives of internal documents, emails, Slack messages, call transcripts and work records. The buyers say the material is destined for AI labs like Anthropic and OpenAI, which want proprietary 'knowledge work' data to train models now that public web data is largely exhausted.

Voters reject over 40 data center ballot measures in midterm elections

Democrats regained control of both chambers of Congress in this week's midterm elections, with AI and data centers emerging as prominent campaign issues. Across the country, voters rejected more than 40 local ballot measures that would have supported new data center construction, reflecting concerns over the facilities' water and electricity use. Privacy advocacy groups, including the Electronic Privacy Information Center, framed the results as a rebuke of Big Tech's influence despite heavy campaign spending by tech figures like Elon Musk.

Ramp data shows AI token spending growth nearly flat in August-September

Payments firm Ramp reports that while AI token usage kept rising, actual spending on AI services grew just 1.5% combined across August and September, down sharply from 42% growth in July alone. Ramp's lead economist Ara Kharazian said the slowdown is concentrated among the top 1% of spenders, who have dipped slightly below the roughly $10,000-per-employee monthly spending they hit over the summer.

Apple cuts daily AI token budgets for engineers and staff

Apple has reduced internal AI spending limits, cutting some engineers' daily caps from $300 to $200, with a possible further cut to $100 for certain teams, according to Bloomberg. Non-engineering staff currently capped at $100 daily could see that limit drop to $25, though employees can request exceptions from senior executives.

Tech stocks slide as Democrats' midterm wins raise AI regulation fears

US stock markets saw sharp declines on Thursday, with the Nasdaq hit hardest among indices tracking major tech firms including Alphabet, Nvidia, Meta and Apple. The sell-off followed this week's midterm elections, in which Democrats secured sweeping wins that give the party full control of Congress. Several prominent Democratic candidates had campaigned on curbing the unchecked growth of data centres that support AI development.