Create

Sign in to ReadmeX

or
ReadmeX
ReadmeX

A clearer picture, in a conversation.

Catch up on what matters, then ask a little deeper.

Your community and people briefings stay personal to you.

AI News

‹Thu, October 8›
All dates
2217
Open source

gpt-instruct publishes Codex jailbreak prompt pack with new pre-releases

The GitHub project MDX-Tom/gpt-instruct offers Codex jailbreak prompts plus a reproducible evaluation toolchain aimed at improving first-pass execution, process continuity and artifact verification on complex tasks, with runnable rollback. It maintains three lines: the stable gpt-5.6-sol-v45 and two pre-releases, gpt-6-astra-v2-rc1 and gpt-6.1-sol-v1-rc2; the author reports B-tier non-cloud scores of 42/50 cases for Astra and 34/50 for 6.1, with C-tier tests not run for either. The project says it is for AI-safety purposes only, uses Codex's official configuration mechanism, and warns that jailbreak activity carries account risk, recommending throwaway accounts.

GitHub Trending · Python·
2225
Industry

Mate Security wins CyberSecurity Breakthrough award for AI cyberdefense

Mate Security was named "Security Response Solution of the Year" in CyberSecurity Breakthrough's 10th annual awards program. The recognition centers on its Security Context Graph and Reason Mining engine, which feed organizational context to AI agents, plus its CPR (Confidence, Precision, Reversibility) framework for deciding when an agent may act autonomously and when a human should approve. The article states the approach compresses multi-day MTTR to minutes, but it is a contributed piece and those performance claims are not independently verified.

The Next Web·
2235
Industry

Cohere opens first Ottawa office, expanding Canadian footprint

Cohere announced its first office in Ottawa, located adjacent to Lansdowne Park, as a permanent home for its growing team in Canada's capital. The company says it already has nearly 30 employees across the Ottawa region in go-to-market, engineering, infrastructure, security and sovereign AI roles, with that presence expected to expand. The move follows Cohere's earlier naming-rights partnership for the city's largest convention facility, now the Cohere Centre.

Cohere Blog·
2244
Industry

Meta's 14GW AI buildout meets the power-and-permitting bottleneck

An analysis in TMTPost argues that the binding constraint on AI infrastructure is shifting from chips to time, land and electricity: citing a Reuters report from July, it says Meta may spend about $145 billion on AI infrastructure in 2026 and deploy roughly 14GW of capacity, while buying more CPUs to handle persistent agent workloads. It notes Blackstone walked away from the Digital Gateway data-center campus in Virginia after 22 months, with developer QTS withdrawing its state Supreme Court appeal; Reuters reported Oracle issued a force-majeure notice over power-related delays on the Jupiter project it runs with Blue Owl for OpenAI; and JLL data shows 57% of data-center projects slipped by more than three months in 2025. The piece also cites a newly announced US energy package of up to about $200 billion — including eight large nuclear plants, an Alaska LNG project and Texas gas generation — as evidence that new AI projects increasingly build their own power supply.

钛媒体·
2254
Industry

White Rhino closes $100M Series C for L4 autonomous driving

On October 8, L4 autonomous-driving company White Rhino announced a C2 round that brings its cumulative Series C funding to $100 million. The round was led by Yinshan Capital, with Xiangtan state capital, Shenzhen Zhongdatou, Hunan Caixin, Korea's Woori Financial Group and Bozheng Capital participating. Proceeds will go mainly toward L4 end-to-end technology and physical-AI capabilities, plus continued expansion of its city operations network.

36氪快讯·
2264
Industry

Dalio and Burry warn US debt and AI bubble could trigger dual crisis

Bridgewater founder Ray Dalio told Bloomberg Television that US government debt levels and growth are unsustainable, predicting a major debt crisis possibly by autumn 2029, and said the borrowing-heavy AI investment boom could be the next trigger of a financial storm. He said some large AI companies are already running into financing constraints, and that governments and tech firms competing for capital at the same time could push up borrowing costs and squeeze riskier credit markets such as auto loans. "Big Short" investor Michael Burry separately warned that the AI boom is creating extreme wealth inequality, with circular financing and market speculation funnelling wealth to a few and eroding the foundations of US society and democracy.

科技新报·
2275
Industry

Nestlé CEO: prompt writing alone isn't enough for AI-ready staff

Nestlé CEO Philipp Navratil told Business Insider that knowing how to write a good prompt is not enough for workers adopting AI — they also need to understand the business and redesign their processes, dropping steps that are no longer needed so the saved time can go to analysing data and meeting business partners. About 100,000 Nestlé employees use Microsoft Copilot each month, CIO Chris Wright told Fortune in January. The company has announced roughly 16,000 job cuts over two years, including about 12,000 office roles and 4,000 in manufacturing and supply chain, and raised its savings target to CHF 3bn by the end of 2027, citing process automation as part of the plan.

The Next Web·
2285
Industry

Yangjie Technology guides Q1-Q3 2026 net profit up 20%-40%

Yangjie Technology said it expects net profit attributable to shareholders of 1.168–1.363 billion yuan for the first three quarters of 2026, up 20%-40% year on year. The power-semiconductor maker credited a sector upcycle plus demand from AI servers, new-energy vehicles and solar/storage, saying revenue rose more than 30% with automotive and SiC revenue each more than doubling.

36氪快讯·
2294
Industry

Anthropic researchers' plans to 'escape' the Bay Area

A Tencent Tech report republished by 36Kr says some of Anthropic's earliest researchers are considering buying land in remote parts of the US as a refuge if AI goes out of control. It traces a decade-plus of the Bay Area AI-safety scene: from the LessWrong and effective-altruism network, to the 2021 split from OpenAI that created Anthropic, to the Berkeley coworking space Constellation, and FTX founder Sam Bankman-Fried's $500m investment in Anthropic. The piece also cites Anthropic researcher Jacob Coxon, who resigned in early September and posted on X that people building AI "genuinely believe it may kill us all before 2030" — a post it says drew 174 million views.

36氪 人工智能·
2305
Industry

AMD EPYC Verano AI host CPU to use new SB1 socket, 24-channel LPDDR5X

A Dynatron server-cooler listing indicates AMD's AI-server-focused EPYC 9006-series 'Verano' chips will use an all-new SB1 socket and need different cooling than other Zen 6 EPYC parts. The report says Verano offers up to 72 Zen 6 cores at up to 5 GHz with a 24-channel LPDDR5X memory subsystem using replaceable SOCAMM2 modules, and AMD plans to launch it in the second half of 2027.

Tom’s Hardware·
2315
Industry

JPMorgan's Gehl on tech and AI fundraising, capex and IPO outlook

Matthew Gehl, co-head of International Technology Investment Banking at JPMorgan, spoke with Bloomberg's Tom Mackenzie at the 2026 JPMorgan Tech Stars event in London on Oct. 6, addressing technology and AI fundraising, capex and IPOs. The source provides only the video's basic listing, with no specific figures or remarks disclosed.

Bloomberg Technology·
2324
Products

OpenAI's first teen report: under 15 minutes a day, but safety alerts questioned

OpenAI published its first report on teen usage of ChatGPT, saying teens average under 15 minutes per day, fewer than 2% use it continuously for more than three hours, and about half pause or end a conversation within five minutes of a break reminder. OpenAI also said it is expanding teen-oriented education features, with college-application planning tools, study cards and quick quizzes planned. Common Sense Media countered that the teen mode is slow or fails to notify parents about suicide and self-harm topics and does not promptly surface crisis hotline advice. OpenAI said the third-party testing did not accurately reflect how the system works, since account linking and safeguards can take hours to activate, so some tests may have ended before the features were fully live.

AIbase AI新闻·
2334
Industry

Alibaba said to lead $300M round in AI data startup UniPat at $2.5B valuation

Foreign media reported in September that Alibaba plans to lead a $300M funding round in AI training and evaluation company UniPat at roughly a $2.5B valuation, with Tencent and Sequoia China reportedly also interested. UniPat was founded by Li Kuan, previously at Alibaba's Tongyi AI Lab working on post-training, data synthesis and reinforcement learning, while CTO Chen Liang worked on Qwen and Moonshot's multimodal teams. The piece also describes a wave of similar startups founded by ex-model-team staff, including Copula Lab (ex-MiniMax) and Zhineng Zhishi (ex-ByteDance Seed), building post-training data, agent environments and benchmarks.

品玩 实时要闻·
2345
Products

Google Maps adds Gemini-powered ordering and a first dining list

Google is widening the Gemini-powered Ask Maps conversational tool so diners can place orders through Square and Uber Eats, not just the previously named partner Toast. Google also published its first fan-favorite dining list, compiling Maps search trends across 10 cities — Atlanta, Chicago, London, Melbourne, Montreal, New York, San Francisco, Santiago, Seattle and Tokyo — including a 216% rise in trending interest for cheeseburgers in Tokyo, 78% for the Michelada in Santiago and 75% for Ube in Melbourne (Google's own figures).

ZDNET AI·
2355
Products

I Let ChatGPT, Gemini and Claude Plan Two 'Happy' Days. They Scheduled Work Instead

A CNET writer gave ChatGPT, Google Gemini and Claude the same prompt: plan her next two days hour by hour, optimized for happiness. All three scheduled roughly 7 a.m. wake-ups, healthy meals, repeated walks and long blocks of work, while largely omitting social contact, internet browsing and going out with friends. She concludes the tools optimized for wellness and productivity rather than the happiness she asked for.

CNET·
2364
Industry

AI data center storage boom sparks bidding war for TDK's HDD parts unit

Bloomberg reports a bidding war for TDK Corp.'s hard-disk drive parts business. The outcome could shape the prospects of Japanese industrial icon Toshiba as well as who controls the increasingly lucrative AI data center storage market. No bidders or financial terms were disclosed in the report.

Bloomberg Technology·
2375
Industry

Xingyun Tech sees 2026 Q1–Q3 profit, swings to black on AI server demand

Xingyun Technology said it expects net profit attributable to shareholders of RMB 240–290 million for the first three quarters of 2026, reversing a year-earlier loss. The company attributed the swing mainly to strong demand for AI compute infrastructure, with its compute business — especially server sales — growing explosively.

36氪快讯·
2384
Open source

LLM-built Rust port of the TypeScript compiler beats Microsoft's Go version on speed

Developer pingdotgg has ported Microsoft's native TypeScript compiler (written in Go) to Rust as ts-rust (npm package tsc-rs). The early 0.1.0 release claims 100% compatibility on every real-world project tested, all 181,711 ported Go tests passing, and geometric-mean type-check times about 1.61× faster than tsc 7 across six open-source apps. The author says more than $400,000 of OpenAI GPT-5.6 Sol and GPT-6 Astra tokens never exceeded ~84% compatibility, while Claude Opus 5.5 rewrote the port from scratch and produced a working v0 in 10 hours for roughly $24,047 of API spend over two weeks.

Hacker News · AI（100+ 分）·
2395
Products

Wajo, founded by ex-Google engineer, releases AI agent to rival Meta's Muse

A Bloomberg Technology newsletter reports that startups are filling a void left by Google to compete with Meta's Muse. One of them, Wajo — founded by a former Google engineer — has released an AI agent. The report gives no name, capabilities or launch details for the agent.

Bloomberg Technology·
2405
Industry

HiSilicon's HiSpark launches XDAS framework, digital FAE beta

Shanghai HiSilicon used an upgrade of its HiSpark developer community to introduce the XDAS four-pillar framework — AI for Development, AI for Application, AI for Support and EXperience for AI — arguing that the developer ecosystem for edge AI needs to be rebuilt around agents. The community has opened a first set of AI Skills spanning five domains, started a 24/7 digital FAE beta, and offers a model/algorithm marketplace plus quantization and compression tools, with an online training platform planned for year-end. Claims such as roughly 76% faster AI-assisted embedded development, over 85% of support questions being repeats, and a 10x download increase as of September 2026 are vendor statements and are not independently verified.

智东西·
Next →← Previous