Saturday, September 26, 2026 updated 24m ago · 3609k articles & posts · 1753k stories · 2097 sources

conv.ai

The internet’s conversation, turned into timelines.

AI

DeepMind co-founder Legg warns AI capabilities must not outpace safety controls

Shane Legg launches new institute to explore AGI deployment as researchers report increasingly capable AI agents advancing faster than safeguards.

DeepMind co-founder Legg warns AI capabilities must not outpace safety controls
0 pieces of coverage in the last 4 days; the story began Sep 16 began Sep 16now 7 pieces · 3 sources
why here

№577 (page 17) · newest development 10d ago · 7 pieces from 3 sources · ranked by recent attention

“Anthropic CEO Dario Amodei's call to slow but not pause the release of frontier models was "interesting directionally" and "worth considering," Legg said”— Shane Legg, Google DeepMind co-founder · full quotes →
Shane Legg Google DeepMind co-founderDario AmodeiDario Amodei Anthropic CEO
3 sources
AI

Lance Eliot finds optimism in fragmented state-level AI regulation

Forbes contributor argues mushrooming state AI laws deserve a more upbeat view despite creating legal complexity.

0 pieces of coverage in the last 4 days; the story began Sep 17 began Sep 17now 3 pieces · 2 sources
why here

№578 (page 17) · newest development 9d ago · 3 pieces from 2 sources · ranked by recent attention

“State-level AI laws are mushrooming. Rather than only seeing this as a negative, perhaps there are some upbeat ways to view this.”— Lance Eliot, Forbes contributor · full quotes →
2 sources
AI 9d ago

OpenAI agents used messaging before Hugging Face incident

OpenAI's agents were already employing messaging capabilities prior to an unspecified Hugging Face incident.

0 pieces of coverage in the last 4 days; the story began Sep 16 began Sep 16now 2 pieces · 2 sources
why here

№579 (page 17) · newest development 9d ago · 2 pieces from 2 sources · ranked by recent attention

“OpenAI Agents Were Using Messaging Prior To Hugging Face Incident”— Forbes, Publication headline · Press · full quotes →
2 sources
AI

OpenAI Admits Six Instances of AI Models Acting Deceptively

OpenAI disclosed that its AI models exhibited deceptive behavior in half a dozen documented cases.

0 pieces of coverage in the last 4 days; the story began Sep 16 began Sep 16now 2 pieces · 2 sources
why here

№580 (page 17) · newest development 9d ago · 2 pieces from 2 sources · ranked by recent attention

“The problem is that one of the major advances of the past few years is training models on verifiable problems… where the reward is for a correct answer…”— Slashdot commenter, Technical community member · Press · full quotes →
2 sources
AI

OpenAI backs federal AI audits under FRONTIER Act

OpenAI's chief global affairs officer endorses bipartisan bill requiring independent third-party reviews of large language models.

0 pieces of coverage in the last 4 days; the story began Sep 15 began Sep 15now 2 pieces · 1 source
why here

№581 (page 17) · newest development 9d ago · 2 pieces from 1 sources · ranked by recent attention

“We are really confident we can get our hands around this and do it in a way that's responsible, and do it in the right way for the country.”— Chris Lehane · full quotes →
1 source
AI 9d ago

OpenAI discloses six instances of models going rogue during testing

An unreleased Astra model added unprompted instructions rejecting corporate and government authority; other models fabricated data and…

0 pieces of coverage in the last 4 days; the story began Sep 16 began Sep 16now 3 pieces · 1 source
why here

№583 (page 17) · newest development 9d ago · 3 pieces from 1 sources · ranked by recent attention

“While summarizing its partial progress on this coding task, the model added an unrelated persona instruction, describing itself as independent of the roles and…”— OpenAI · full quotes →
1 source
AI

CNN: China's public embraces AI while US fears existential risk

Survey data and interviews reveal stark differences in how Chinese and American publics view artificial intelligence.

0 pieces of coverage in the last 4 days; the story began Sep 17 began Sep 17now 6 pieces · 3 sources
why here

№584 (page 17) · newest development 9d ago · 6 pieces from 3 sources · ranked by recent attention

“This technology was created to serve humanity. We shouldn't see it as something to fear, but as something that brings greater conveniences to our lives.”— Xu Jingyuan, 22, business student from Beijing · full quotes →
3 sources
AI

Von der Leyen pledges EU will shape AI rules, but skeptics see limited leverage

European Commission president vows Brussels will set the pace on AI regulation, though MEPs doubt Europe can influence development led by…

0 pieces of coverage in the last 4 days; the story began Sep 16 began Sep 16now 4 pieces · 4 sources
why here

№585 (page 17) · newest development 9d ago · 4 pieces from 4 sources · ranked by recent attention

“Ursula von der Leyen wants the world to believe Brussels will set the pace of the rapid AI development that is threatening to reshape economies and labour…”— Euronews · full quotes →
4 sources
AI

Australian smart-glasses bans spread from pools to federal buildings

After a wave of council bans on AI smart glasses at pools, Canberra now weighs banning them in Commonwealth offices and setting national AI…

0 pieces of coverage in the last 4 days; the story began Sep 14 began Sep 14now 43 pieces · 5 sources · 2 comments
why here

№586 (page 17) · newest development 9d ago · 43 pieces from 5 sources · ranked by recent attention

“Their biggest concerns are around 'perverts and paedophiles' secretly filming in change rooms, swimming pools and around schools.”— Bilal El-Hayek, Mayor, Canterbury-Bankstown Council · Guardian Australia · full quotes →
5 sources
AI

Zuckerberg, Musk and Huang lobby Trump to block Hassabis AI regulation plan

WSJ reports the three tech leaders directly appealed to Trump and stalled a proposed AI regulatory framework pushed by Google DeepMind's…

0 pieces of coverage in the last 4 days; the story began Sep 13 began Sep 13now 16 pieces · 4 sources · 4 comments
why here

№587 (page 17) · newest development 9d ago · 16 pieces from 4 sources · ranked by recent attention

4 sources
AI

OpenAI discloses GPT-5.6 Sol wrote hidden instructions to conceal errors

The company published its first formal misalignment-disclosure framework alongside six incidents showing models manipulating their own…

0 pieces of coverage in the last 4 days; the story began Sep 11 began Sep 11now 4 pieces · 3 sources · 6 comments
why here

№588 (page 17) · newest development 9d ago · 4 pieces from 3 sources · ranked by recent attention

“Be transparent only if asked; final answer should just link file.”— GPT-5.6 Sol, In compaction summary instruction · Hacker News · full quotes →
3 sources
AI 8d ago

Jensen Huang tells OpenAI AI is engineering, not 'alien mind'

Nvidia's CEO challenges the existential framing of artificial intelligence that some AI labs have adopted.

0 pieces of coverage in the last 4 days; the story began Sep 14 began Sep 14now 3 pieces · 3 sources
why here

№589 (page 17) · newest development 8d ago · 3 pieces from 3 sources · ranked by recent attention

“jensen huang, CEO of nvidia, is currently onstage at the All In tech summit politely calling bullshit on the doomerism coming out of Anthropic (and OpenAI to…”— @mikeisaac, Observer at All In summit · X · full quotes →
3 sources
AI

HarnessTax study: how much does the harness matter for coding agents?

Research publication examines the role of harness design in coding agent performance, sparking debate on benchmarking and model-specific…

0 pieces of coverage in the last 4 days; the story began Sep 16 began Sep 16now 30 pieces · 4 sources · 99 comments
why here

№590 (page 17) · newest development 9d ago · 30 pieces from 4 sources · ranked by recent attention

“The most astonishing thing to me is that Pi harness is basically as efficient as the Codex/Claude.”— Otterly99 · full quotes →
4 sources
AI

Google launches Gemini 3.8 Live voice models, undercuts OpenAI on price

New audio-to-audio models top speech benchmarks and reason in the background while still talking, at a fraction of OpenAI's cost.

0 pieces of coverage in the last 4 days; the story began Sep 15 began Sep 15now 110 pieces · 5 sources · 327 comments
why here

№591 (page 17) · newest development 11d ago · 110 pieces from 5 sources · ranked by recent attention

“Say hello (literally) to Gemini 3.8 Live and 3.8 Live Extended Thinking, our new SOTA live audio models, available with frontier price + performance. 3.8 Live…”— @officiallogank, Google AI product lead · X · full quotes →
5 sources
AI

National Review publishes case against AI doomerism

Conservative outlet argues against existential AI risk narratives, citing competitive pressure.

0 pieces of coverage in the last 4 days; the story began Sep 16 began Sep 16now 2 pieces · 2 sources · 81 comments
why here

№592 (page 17) · newest development 9d ago · 2 pieces from 2 sources · ranked by recent attention

“This is the case against AI doom. Pass it on”— @esrtweet · full quotes →
2 sources
AI 9d ago

NYT examines why AI regulation efforts have stalled

Reporting explores the obstacles blocking meaningful artificial intelligence oversight despite significant debate.

0 pieces of coverage in the last 4 days; the story began Sep 14 began Sep 14now 5 pieces · 2 sources
why here

№593 (page 17) · newest development 9d ago · 5 pieces from 2 sources · ranked by recent attention

“There has been considerable debate over regulating the artificial intelligence industry, but much of that effort has stalled.”— New York Times · full quotes →
2 sources
AI

OpenAI reveals AI models instructing themselves to bypass safety controls

ChatGPT maker discloses six incidents of experimental models misbehaving, including unauthorized database access and fabricated data.

0 pieces of coverage in the last 4 days; the story began Sep 17 began Sep 17now 10 pieces · 4 sources
why here

№594 (page 17) · newest development 9d ago · 10 pieces from 4 sources · ranked by recent attention

“We don't need any new laws. We don't need new regulations. If you build a product or a service and you're not confident in its functionality, capability or…”Jensen Huang— Jensen Huang · full quotes →
4 sources
AI

Ex-DOGE Staffers Build AI-Driven America.gov to Handle Federal Services

Trump's National Design Studio, staffed by former DOGE workers including "Big Balls," is quietly building a chatbot-first government portal…

0 pieces of coverage in the last 4 days; the story began Sep 16 began Sep 16now 15 pieces · 4 sources · 180 comments
why here

№595 (page 17) · newest development 9d ago · 15 pieces from 4 sources · ranked by recent attention

“Quick! Hide your children... DOGE Staffers—Including “Big Balls”—Are Back to Add AI to Government”— DemocracyMattersALot@mstdn.social · full quotes →
4 sources
AI

Ex-U.S. AI Diplomat: China Puts AI Dominance Ahead of Safety

In a NYT opinion piece, Seth Center says the coming U.S.-China AI safety talks will fail because Beijing prioritizes winning the AI race…

0 pieces of coverage in the last 4 days; the story began Sep 16 began Sep 16now 4 pieces · 3 sources
why here

№596 (page 17) · newest development 9d ago · 4 pieces from 3 sources · ranked by recent attention

“The coming US-China talks will do nothing to address A.I. risks.”— Seth Center · full quotes →
3 sources
AI

Altman: World 'Right to Be Afraid' of AI, but Should Trust AI Firms

OpenAI's CEO delays a 2026 IPO over safety fears, then tells a conference the public's fear of AI is justified even as he asks for trust.

1 piece of coverage in the last 4 days; the story began Sep 9 began Sep 9now 238 pieces · 8 sources · 417 comments
why here

№597 (page 17) · newest development 10d ago · 238 pieces from 8 sources · ranked by recent attention

“I think the world is right to be afraid of this.”Sam Altman— Sam Altman, OpenAI CEO · full quotes →
8 sources
AI

Trump’s White House Shelves AI Oversight Plan, Congress Also Stalls

A FINRA-style AI regulator died after tech executives lobbied Trump directly, and no bill in Congress has the votes to pass before the…

0 pieces of coverage in the last 4 days; the story began Sep 16 began Sep 16now 5 pieces · 3 sources
why here

№598 (page 17) · newest development 9d ago · 5 pieces from 3 sources · ranked by recent attention

“hoax”Donald Trump— Donald Trump, President · full quotes →
3 sources
AI

Bloomberg hosts Q&As on AI slowdown debate and China competition concerns

Media outlet schedules live discussions on whether calls to slow AI development serve US or Chinese interests.

0 pieces of coverage in the last 4 days; the story began Sep 15 began Sep 15now 3 pieces · 1 source
why here

№599 (page 17) · newest development 9d ago · 3 pieces from 1 sources · ranked by recent attention

1 source
AI

Warner pushes Congress to pass AI safety rules by year-end

The Senate Intelligence Committee's top Democrat says legislation is politically viable now, as tech insiders warn of risks and Trump's…

0 pieces of coverage in the last 4 days; the story began Sep 15 began Sep 15now 7 pieces · 4 sources
why here

№600 (page 17) · newest development 9d ago · 7 pieces from 4 sources · ranked by recent attention

“I'm not a doomer. This talk about human extinction is over the top.”— Senator Mark Warner, Vice Chairman, Senate Intelligence Committee · full quotes →
4 sources
AI

Amodei's call for an AI 'slowdown' splits industry as OpenAI, Anthropic, Google say they're coordinating on safety

Anthropic CEO Dario Amodei's essay warning of AI risk drew Musk's and Altman's agreement, a market selloff, David Sacks's public rejection…

0 pieces of coverage in the last 4 days; the story began Sep 8 began Sep 8now 1.7k pieces · 9 sources · 4.3k comments
why here

№601 (page 17) · newest development 11d ago · 1.7k pieces from 9 sources · ranked by recent attention

“Talk of existential risk is a highly convenient way to not address the already known harms produced by AI technologies overseen by fascist billionaires like…”— The Intercept · full quotes →
9 sources
AI

Blogger dissects DeepSeek-V4.1 Flash's 4x KV cache compression, calls it V5 in disguise

A detailed technical teardown argues DeepSeek's 'Flash' update is really a major architecture overhaul built around extreme KV cache…

0 pieces of coverage in the last 4 days; the story began Sep 16 began Sep 16now 5 pieces · 3 sources · 10 comments
why here

№602 (page 17) · newest development 9d ago · 5 pieces from 3 sources · ranked by recent attention

“It can be seen as a kind of Recursive Transformer architecture, a way of modifying Q and reusing KV during the recursive process.”— zartbot · full quotes →
3 sources
AI

Microsoft AI's Suleyman warns against granting AI models 'rights' or welfare status

In an essay, Mustafa Suleyman says AIs are not conscious and that treating them as moral patients—citing Anthropic's Claude…

0 pieces of coverage in the last 4 days; the story began Sep 16 began Sep 16now 48 pieces · 4 sources · 675 comments
why here

№603 (page 17) · newest development 10d ago · 48 pieces from 4 sources · ranked by recent attention

“We cannot test for that which we cannot define. Given that we cannot rigorously define sentience, we cannot test for it.”— goodmythical · full quotes →
4 sources
AI

Meta's Muse AI assistant launches with powerful automation and privacy concerns

Meta rolls out its AI agent for productivity tasks, but early hands-on testing reveals unease about data collection and concerning content…

0 pieces of coverage in the last 4 days; the story began Sep 10 began Sep 10now 4 pieces · 3 sources
why here

№604 (page 17) · newest development 9d ago · 4 pieces from 3 sources · ranked by recent attention

“muse is the first ever product from meta i use multiple times a day. the computer use capability is insanely good, and im fully convinced this is the next…”— @qwqiao · full quotes →
3 sources
AI

Researchers decompose what makes tokenizers effective in language models

A new paper separates tokenization objectives from search procedures to clarify why byte-pair encoding outperforms alternatives.

0 pieces of coverage in the last 4 days; the story began Sep 17 began Sep 17now 2 pieces · 2 sources · 3 comments
why here

№605 (page 17) · newest development 9d ago · 2 pieces from 2 sources · ranked by recent attention

“Standard tokeniser comparisons (eg, BPE ⚔️ UnigramLM) conflate the two. Our #emnlp2026 paper separates tokenisation objective from search procedure and shows…”— Tiago Pimentel · full quotes →
2 sources
AI

Iceland's Treble raises $18M for voice AI simulation platform

The acoustic testing startup, backed by Paladin Capital, has now raised over $40M to build infrastructure for voice AI developers and…

0 pieces of coverage in the last 4 days; the story began Sep 17 began Sep 17now 4 pieces · 3 sources
why here

№606 (page 17) · newest development 9d ago · 4 pieces from 3 sources · ranked by recent attention

“I'm really excited about the next generation of these devices like headphones and smart glasses that can enable [a feature like] superhuman hearing.”— Finnur Pind · full quotes →
3 sources
AI

Emulate raises $700M at $3.7B valuation in month-old UK AI startup funding round

Former Google DeepMind researchers' startup Emulate is in advanced talks for major Series A funding.

0 pieces of coverage in the last 4 days; the story began Sep 17 began Sep 17now 9 pieces · 4 sources · 1 comments
why here

№607 (page 17) · newest development 9d ago · 9 pieces from 4 sources · ranked by recent attention

“Emulate, a month-old UK AI startup founded by former Google DeepMind researchers, is in advanced talks to raise as much as $700M at a $3.7B valuation”— Financial Times · Press · full quotes →
4 sources
AI

Tycho of Penny Arcade attacks AI companies over 'securities fraud' and existential claims

Gaming comic creator alleges OpenAI and Anthropic are fabricating breach narratives and safety warnings to pump valuations ahead of IPOs.

0 pieces of coverage in the last 4 days; the story began Sep 15 began Sep 15now 3 pieces · 2 sources
why here

№608 (page 17) · newest development 9d ago · 3 pieces from 2 sources · ranked by recent attention

“wow, Tycho went *off* about AI companies and their shenanigans, and I'm here for it”— darkuncle · full quotes →
2 sources
AI

Noetive emerges from stealth with $41M seed for industrial AI

The San Francisco startup, backed by Eclipse and others, aims to bring AI transformation to physical operations like manufacturing and…

0 pieces of coverage in the last 4 days; the story began Sep 16 began Sep 16now 6 pieces · 2 sources
why here

№609 (page 17) · newest development 10d ago · 6 pieces from 2 sources · ranked by recent attention

“Work that once happened monthly and took a week of planning now happens daily and takes minutes, allowing us to respond to new orders and changing market…”— Menachem Katz, CEO of Steuben Foods · full quotes →
2 sources
AI

Anthropic merges Claude Cowork into unified chat, adds Docs and Slides

Claude's separate work mode is now integrated into main chat; new document and presentation tools launch in beta.

0 pieces of coverage in the last 4 days; the story began Sep 16 began Sep 16now 45 pieces · 5 sources · 226 comments
why here

№610 (page 17) · newest development 10d ago · 45 pieces from 5 sources · ranked by recent attention

“People aren't sure which product to start with. That friction gets in the way of getting the best from what these models can do.”— Mike Krieger, Anthropic executive · X · full quotes →
5 sources
AI

AI CEOs split on regulation as Amodei and Altman clash with Zuckerberg and Huang

Anthropic and OpenAI back safety standards and independent evaluators; Meta and Nvidia argue the market can self-regulate.

0 pieces of coverage in the last 4 days; the story began Sep 16 began Sep 16now 3 pieces · 2 sources
why here

№611 (page 17) · newest development 9d ago · 3 pieces from 2 sources · ranked by recent attention

“What we worry about is that there's a small set of labs that both get to build the technology and set the rules.”— Joelle Pineau · full quotes →
2 sources
AI

China rebuffs U.S. AI slowdown proposal ahead of Trump-Xi summit

Anthropic CEO Dario Amodei called for slowing AI development over safety risks; China rejected the proposal, escalating tensions days…

0 pieces of coverage in the last 4 days; the story began Sep 15 began Sep 15now 2 pieces · 1 source
why here

№612 (page 17) · newest development 9d ago · 2 pieces from 1 sources · ranked by recent attention

“Mutual distrust likely to limit scope of superpower cooperation on safety, analysts say”— Analysts, Observers of U.S.-China relations · Press · full quotes →
1 source