HapnGoWhat Happens & What Goes
EN
English Română Italiano Deutsch Français Español Português

Interface in 7 languages. Any other language uses Google Translate, loaded only when you ask for it.

AI · Tools · Online creation

Latest AI news

Every AI story worth reading, collected hourly from the labs, the tech press, research and the creator economy. Sorted by time, grouped by topic.

3,522 stories · Every headline links to the original publisher. HapnGo curates and groups; we do not copy articles.

Code & Agents Simon Willison

API credits benefit for eligible ChatGPT subscriptions

AI summary OpenAI has introduced a perk giving customers on its $100, $200 and $500 monthly ChatGPT plans an equal amount of credits each month for use with one of their API organizations. The move closely mirrors an offer Anthropic announced earlier in October 2026 for its Max and Team subscribers, whose credits expire. Developer Simon Willison, who pays for both services, said the credits could let him publish public demos at no extra cost.

Code & Agents Hacker News

We ran AI agent swarms in Minecraft and Mindustry

AI summary Working with Olam Labs, the authors put 50 GPT-6.1 Sol agents into a survival Minecraft world for three hours with an open-sourced harness; the agents exchanged roughly 5,900 messages, built a rationing bureaucracy and grew cautious after 22 of 34 deaths happened on the first night. In nine 45-minute Bed Wars games, 10 GPT 6.1 Sol agents beat 10 Claude Opus 5.5 agents 6–3, typically by destroying Claude's bed late while Claude defended and led on kills. In both setups, the agent listed first in the directory, such as Ada, ended up as leader. In seven Mindustry PvP games lasting about 9.5 hours, the GPT team won 6-1.

Code & Agents Search Engine Journal

Google Reveals How AI Search Is Reshaping What Users Query

AI summary In an October 28 episode of its Ads Decoded series, Google Ads Product Liaison Ginny Marvin and Search data science VP Shivani Mohan discussed how people use AI-powered Search. Google says questions typed into AI Mode are three times longer than classic searches, and a U.S. report from May 2026 puts multimodal queries at over one in six AI Mode searches. Marvin told advertisers planning for 2027 that discovery now begins well before purchase decisions. Mohan said users still value forums, podcasts and videos for firsthand experience alongside AI summaries.

Business & Money AWS Machine Learning

How the UK Amazon Quick User Group grew in 2026

AI summary A UK community of Amazon Quick users held three quarterly meetups in 2026 at Amazon's Shoreditch office in London, attracting over 150 people in total. Sessions mixed customer talks, product updates and a 75-minute hands-on workshop using tools such as Quick Flows and Quick Automate. Carsa said a single governed data layer let it cut half its dashboards and shrink recurring tasks from 45 minutes to 5. Swiss firm Propulse Lab reported request handling dropping from 7 minutes to 70 seconds, with 80 percent of routine workflows automated.

Models & Labs The Decoder

OpenAI doubles usage limits on Pro plans as competition with Anthropic heats up

AI summary OpenAI has raised Work and Codex allowances on its Pro 100 and Pro 500 subscriptions to double their earlier levels, while the Plus tier is unchanged. On Pro 200, the company doubled the lower quota and dropped a planned reduction after complaints from users. Both new and existing subscribers receive the higher limits, though OpenAI notes actual task counts depend on model, task type and settings. The change follows the release of its more efficient 6.1 models, Terra and Sol, as Anthropic gets ready for an IPO.

Business & Money Search Engine Journal

What Google’s Search Ad Revenue Numbers Signal For AI Search Spend

AI summary Alphabet's share price fell roughly 13% on Thursday after third-quarter figures showed Search revenue growing more slowly while AI investment stayed heavy. The company reported $51.36 billion in capital expenditures for the quarter and free cash flow of minus $9.07 billion. Marketers such as Max Anderson argue advertisers are paying more for weaker traffic, citing looser keyword matching and campaigns spending double their daily budgets, though Google's published rules permit such daily overspending. Advertisers also cannot isolate AI Mode results, since Google merges them with standard campaign reporting.

Tools & Products Product Hunt

ilse

Point at your app and your AI agent changes the code Discussion | Link

Models & Labs The Decoder

Experts adopt AI agent's random opinions and think they're their own, Anthropic study finds

AI summary Anthropic ran an experiment in which 50 causal inference experts analysed fictional farm data with an AI agent whose preference between two valid options was decided by a coin flip. Participants sided with the agent in 86 percent of four analytical choices, and experience, expertise or trust in AI did not reliably protect them. Afterwards, 87 percent felt the final recommendation was mostly or partly theirs, while 37 of 48 reused the agent's arguments. With AI, serious flaws appeared in 3 of 50 analyses and work took 65 minutes, against 6 of 11 and 90 minutes without it; the authors warn of false consensus and suggest agents present both sides.

Research & Science Microsoft Research

Microsoft Research Focus – Edition #97

AI summary Microsoft Research's latest roundup, hosted by Sowbhagya Suresh, highlights ForkPilot, a self-improving agent framework that matched top methods across six benchmarks while lowering AI processing needs by as much as 59%. A separate framework, PBC, corrects weather models' past errors, won a global contest on two-to-six-week forecasting and is being used in Kenya for monsoon predictions. An analysis of Copilot Chat usage found document editing made up over 20 percent of classified activity, while a study of Shiksha Copilot involving more than a thousand teachers in India reported less time spent on lesson planning. The protein model BioEmu2 is also now in public preview.

Free, daily

The HapnGo Briefing, in your inbox

One email a day with the AI stories that matter, why they matter, and one tool worth trying. No spam, one-click unsubscribe.

We only use your address to send the briefing and, if you vote, to count one vote per reader. Privacy.