Collection window: articles published between 2026-11-01 20:47 UTC and 2026-11-02 20:47 UTC (last 24 hours). All timestamps converted to UTC. Stories from the previous brief are excluded.
1. OpenAI merges ChatGPT’s Chat and Work tabs on desktop — and puts ChatGPT and Codex on one usage pool
In an email to users, OpenAI said it is combining the Chat and Work tabs of the ChatGPT desktop app into a single experience. From one screen, users can bring in files and plugins and have ChatGPT carry an idea through to a finished website, presentation or web task while they review its work. The merged view will reach select users first for feedback before a wider rollout, and Codex stays a separate space in the app. ChatGPT and Codex will now draw from the same token allowance, so some heavy users may hit usage resets more often. Anthropic merged Claude’s chat and Cowork functions into one screen in September.
Reference: TechCrunch — 2026-11-02 13:57 UTC
2. Abu Dhabi signs four OpenAI deals; ADNOC becomes the first energy company on OpenAI’s Frontier platform
Abu Dhabi entities and OpenAI signed four agreements at ADIPEC, witnessed by Crown Prince Sheikh Khaled bin Mohamed and Sam Altman. State energy giant ADNOC becomes the first energy firm to deploy OpenAI’s Frontier platform in a multi-year partnership with G42 subsidiary Inception42, with inference running inside the UAE. The Department of Government Enablement launched a Machine-Readable Government initiative that converts policy into AI-parsable form (piloted on food imports); Mubadala signed on to apply AI to investment analysis and risk assessment; and Inception42 agreed a five-year Frontier rollout across government, enterprise and education. No contract values were disclosed; the deals sit under the Dh13 billion ($3.5 billion) AI-native government push in Abu Dhabi’s 2025–2027 Digital Strategy.
Reference: The National / AI Weekly summary — 2026-11-02 06:44 UTC
3. SemiAnalysis launches GDPx agentic benchmark: the best model scores just 12%, and no model leads everywhere
SemiAnalysis released GDPx, an interactive agentic benchmark spanning 11 domains including chip design, cybersecurity, biology, finance, legal, quant trading, recursive self-improvement and materials science. The top model scores about 12%, and every model scored zero on finance and legal under binary grading. Opus 5.5 leads quant trading, hardware and chip design; Gemini 4 Carbon leads biology and software engineering; GPT-6 Astra leads recursive self-improvement and materials science. The authors say safety-classifier refusals dragged down Opus 5.5 and GPT-6 Astra on the cyber, bio and alignment splits, and contrast GDPx with OpenAI’s GDPval by using realistic environments such as a deal room from a real $100B+ private-equity transaction.
Reference: SemiAnalysis — 2026-11-02 (last modified 02:00 UTC) / AI Weekly summary — 2026-11-02 03:32 UTC
4. Linux 7.4-rc1 lands the biggest merge window ever; Torvalds calls AI fixes arriving at merge time a “much healthier model”
Linus Torvalds released the first release candidate of Linux 7.4, saying the merge window brought in nearly 20,000 commits (19,671) versus the usual ~15,000. He said he is much happier to see the flood of AI-found fixes land during the merge window rather than piling up as extra fixes in the -rc stabilization period, and hopes AI fixes become part of normal development. He also noted that 7.4 is expected to be the next LTS release and that AMD GPU register definitions make up about a third of the changes. By AI Weekly’s count, commits declaring AI assistance rose about 2.4x, and Torvalds added that AI-suggested patches are often rewritten.
Reference: The Register — 2026-11-01 23:11 UTC
5. OpenAI readies a post-midterm PR blitz and may rebrand data centers as “AI compute campuses”
According to Axios, OpenAI is preparing a major PR campaign for after Tuesday’s US midterm elections. Chief global affairs officer Chris Lehane called the midterms “a bit of a movie trailer” for 2028, when he expects AI to be the central issue of the presidential race. The playbook includes AI “speed limits,” showcasing how AI empowers people day to day, pushing the US to help set global standards, and renaming data centers “AI compute campuses.” Amid growing local opposition to data centers, Amazon and Microsoft have said they will stop using NDAs in negotiations with local governments.
Reference: TechCrunch — 2026-11-01 21:55 UTC
6. South Korea’s 2027 budget: ₩21.3 trillion for chips, data centers and physical AI, plus ₩4.6 trillion for a sovereign AI model
In a speech to the National Assembly, President Lee Jae Myung asked for bipartisan support for a ₩820.9 trillion (about $614.5 billion) 2027 budget, up more than 10%, highlighting investment in AI and other growth industries. The government earmarked ₩21.3 trillion for its “three megaprojects” covering a semiconductor manufacturing hub, data centers, physical AI and AI, and a separate ₩4.6 trillion to acquire a sovereign AI model aimed at world-leading performance. Another ₩13.8 trillion goes to 10 strategic technology sectors, alongside a ₩160 trillion “Future Fund” for younger generations, growth engines and regional development.
Reference: The Korea Herald — 2026-11-02 01:48 UTC (10:48 KST)
7. LG CNS completes a 30MW hyperscale AI data center in Jakarta — the first built by a Korean firm for an overseas client
LG CNS finished a 30-megawatt hyperscale AI data center in Jakarta, Indonesia, commissioned by KMG, a joint venture of Indonesia’s Sinar Mas Group and Korea Investment Real Asset Management. It is the first AI data center a South Korean company has built for an overseas client, designed for next-generation GPU workloads and prepared to scale to up to 400MW. LG CNS plans to expand into Malaysia, Thailand, Vietnam and Japan and to add subscription GPU services, liquid-cooling infrastructure and AI factory projects.
Reference: The Korea Herald — 2026-11-02 02:37 UTC (11:37 KST)
Today’s Takeaway
- OpenAI is pushing on three fronts at once — product (Chat/Work merge, one usage pool with Codex), foreign governments (four Abu Dhabi deals) and domestic politics (a post-midterm PR push) — to secure both agentic work tools and public legitimacy for its infrastructure.
- GDPx’s 12% ceiling and the surge of AI fixes in Linux 7.4 show AI moving deep into real work while still far from finishing expert tasks on its own. Korea is accelerating in the AI infrastructure race through its budget and LG CNS’s first overseas AI data center.
Leave a Reply