AINews Portal

News

50 items (filtered)

The raw AI news stream — everything we ingest across RSS, Hacker News, arXiv, GitHub, Bluesky, YouTube and more, newest first.

SourceAgeTitleSummaryStatus
arxiv
arxiv-ai
2moAn Exponential Deterministic--Randomized Gap in ERM-Oracle Complexity for Thresholds on an Unknown OrderAttias, Hanneke and Ramaswami (NeurIPS 2025) asked whether randomization provably reduces the oracle calls needed for online learning…
↗Not briefed
bluesky
bluesky-ai
2moMajor update: Three Anthropic employees warn that AI could lead to human extinction within the next decade. Trump refuses regulation. DHS…Major update: Three Anthropic employees warn that AI could lead to human extinction within the next decade. Trump refuses regulation.…
↗Not briefed
arxiv
arxiv-ai
2moAre You Learning Biological Signal or Shortcuts? Auditing and Mitigating Bias in Protein-Protein Interaction DatasetsProtein-protein interaction (PPI) databases do not faithfully reflect biological realities. Instead, they are influenced by study and…
↗Not briefed
arxiv
arxiv-ai
2moWho Argues What? Joint Argument-Entity Detection and Classification in Political Debates🔧 ReplicatePolitical debates are often analyzed through Argument Mining (AM) to investigate the key arguments that drive them. However, political…
↗Not briefed
hn
hn-ai
2moAnthropic insiders warn AI could kill all humans—
↗Not briefed
bluesky
bluesky-ai
2moI still think the biggest threats from software automation have very mundane human origins, and the claims of looming superintelligence bot…I still think the biggest threats from software automation have very mundane human origins, and the claims of looming superintelligenc…
↗Not briefed
bluesky
bluesky-ai
2mowhile hacking software automation will certainly be a real problem the anthropic engineer is a kid who is pushing the idea that AI is on th…while hacking software automation will certainly be a real problem the anthropic engineer is a kid who is pushing the idea that AI is…
↗Not briefed
bluesky
bluesky-ai
2moAn Anthropic researcher just quit, warning they're racing to superintelligence. An employee still there agreed, and put the odds of AI kil…An Anthropic researcher just quit, warning they're racing to superintelligence. An employee still there agreed, and put the odds of A…
↗Not briefed
bluesky
bluesky-ai
2mothe reason i dont like the bernie plan of the government owning half of Open and Anthropic is specifically that it is committing the govern…the reason i dont like the bernie plan of the government owning half of Open and Anthropic is specifically that it is committing the g…
↗Not briefed
bluesky
bluesky-ai
2moThe fact that ~anybody with a strong AI model and tokens can reprouce/redo/steal ~any idea from descriptions is a big argument against tran…The fact that ~anybody with a strong AI model and tokens can reprouce/redo/steal ~any idea from descriptions is a big argument against…
↗Not briefed
bluesky
bluesky-ai
2moCombien de temps avant qu'un LLM dopé au Nouveau Testament soit adoré par une hérésie en tant que retour du Christ, un genre de YeshuAI tra…Combien de temps avant qu'un LLM dopé au Nouveau Testament soit adoré par une hérésie en tant que retour du Christ, un genre de YeshuA…
↗Not briefed
bluesky
bluesky-ai
2moSo OpenAi stole the math, the same way they steal everythingSo OpenAi stole the math, the same way they steal everything
↗Not briefed
bluesky
bluesky-ai
2moTeniendo en cuenta que el señor CEO de Anthropic dijo estando con drogas hasta arriba que mataba gente y eso era normal, me lo creo.Teniendo en cuenta que el señor CEO de Anthropic dijo estando con drogas hasta arriba que mataba gente y eso era normal, me lo creo.
↗Not briefed
bluesky
bluesky-ai
2mowhoa but I thought Anthropic were The Good Ones™whoa but I thought Anthropic were The Good Ones™
↗Not briefed
hn
hn-ai
2moMy LLM eval cried wolf. Here's what I measured—
↗Not briefed
hn
hn-ai
2moA remote island, these GPS hackers are preparing for an invisible war—
↗Not briefed
bluesky
bluesky-ai
2moAnthropic executive here predicts that perhaps 10 percent of the world population could be wiped out by rogue AI that he and other executiv…Anthropic executive here predicts that perhaps 10 percent of the world population could be wiped out by rogue AI that he and other exe…
↗Not briefed
arxiv
arxiv-ai
2moCan AI Agents Deliver Verifiable Network-Wide Outcomes Across Authority Boundaries?AI agents are increasingly involved in network automation, where they can initiate configuration changes through mediated operational…
↗Not briefed
rss
techmeme
2moSources: Iran's central bank has quietly relaxed foreign currency controls and encouraged traders to repatriate funds via cryptocurrencies to bypass sanctions (Najmeh Bozorgmehr/Financial Times)Najmeh Bozorgmehr / Financial Times: Sources: Iran's central bank has quietly relaxed foreign currency controls and encouraged traders…
↗Not briefed
bluesky
bluesky-ai
2moAn Anthropic researcher is quitting the artificial-intelligence industry over fears that the lab and its competitors are racing to build sy…An Anthropic researcher is quitting the artificial-intelligence industry over fears that the lab and its competitors are racing to bui…
↗Not briefed
bluesky
bluesky-ai
2moFormer Anthropic researcher quits, warning: “The people building AI earnestly believe that it could kill us all by the end of the decade” …Former Anthropic researcher quits, warning: “The people building AI earnestly believe that it could kill us all by the end of the deca…
↗Not briefed
bluesky
bluesky-ai
2moEl tema del día es que Evan Hubinger que es un director en Anthropic ubica la posibilidad de que la IA nos mate a todos en el 10%.El tema del día es que Evan Hubinger que es un director en Anthropic ubica la posibilidad de que la IA nos mate a todos en el 10%.
↗Not briefed
bluesky
bluesky-ai
2moi spend a lot of my time each day in Microsoft Fucking Sharepoint so i'm not exactly unsympathetic when guys overcalibrate from a few good …i spend a lot of my time each day in Microsoft Fucking Sharepoint so i'm not exactly unsympathetic when guys overcalibrate from a few…
↗Not briefed
bluesky
bluesky-ai
2moAn anthropic exec just put it out there that he thinks his company and others like it could murder more than 10 percent of humanity with it…An anthropic exec just put it out there that he thinks his company and others like it could murder more than 10 percent of humanity wi…
↗Not briefed
bluesky
bluesky-ai
2moYour 'bonus moment of doom' for today: "The people building AI earnestly believe that it could kill us all by the end of the decade... No …Your 'bonus moment of doom' for today: "The people building AI earnestly believe that it could kill us all by the end of the decade..…
↗Not briefed
hn
hn-ai
2moShow HN: Castforge – run Claude Code, Codex and Gemini as one dev team—
↗Not briefed
hn
hn-ai
2moShow HN: Mnemiq – Open-source text-to-SQL tuned to your own database—
↗Not briefed
bluesky
bluesky-ai
2moI've had a hefty ol' rant about the ridiculous "AI WILL DESTROY US ALL!" lies coming from the companies making the genAI. kotaku.com/anthr…I've had a hefty ol' rant about the ridiculous "AI WILL DESTROY US ALL!" lies coming from the companies making the genAI. kotaku.com/…
↗Not briefed
hn
hn-ai
2moUpyo 0.6.0: MIME composition, streaming attachments, and calendar invitations—
↗Not briefed
bluesky
bluesky-ai
2moAnthropic Researcher Declares AI Could Wipe Out Humanity In 10 Years (It Won't) https://kotaku.com/anthropic-ai-genai-claude-openai-opinion…Anthropic Researcher Declares AI Could Wipe Out Humanity In 10 Years (It Won't) https://kotaku.com/anthropic-ai-genai-claude-openai-op…
↗Not briefed
bluesky
bluesky-ai
2moHe added that “people building AI earnestly believe that it could kill us all by the end of the decade.” Evan Hubinger, an alignment scien…He added that “people building AI earnestly believe that it could kill us all by the end of the decade.” Evan Hubinger, an alignment…
↗Not briefed
bluesky
bluesky-ai
2moYou can’t just post Tweets by apocalypse-predicting Anthropic employees without posting the comments, folks. That’s where the juice is!!!You can’t just post Tweets by apocalypse-predicting Anthropic employees without posting the comments, folks. That’s where the juice is…
↗Not briefed
arxiv
arxiv-ai
2moBeyond Surface Imitation: Contrastive Modeling for Reasoning Path Alignment in Multimodal In-Context LearningIn-context learning (ICL) is widely used in multimodal large language models (MLLMs) and achieves strong performance across a wide ran…
↗Not briefed
bluesky
bluesky-ai
2moLLM: this code is special it's impossible to test. I need to exclude it from the coverage reports. Me: No, it's not. LLM: You're right, I…LLM: this code is special it's impossible to test. I need to exclude it from the coverage reports. Me: No, it's not. LLM: You're rig…
↗Not briefed
bluesky
bluesky-ai
2moNot surprising that one of the first things Anthropic is doing with its AI model is spying on and tracking activists who oppose Anthropic. …Not surprising that one of the first things Anthropic is doing with its AI model is spying on and tracking activists who oppose Anthro…
↗Not briefed
bluesky
bluesky-ai
2mo"But it's too detailed!" "The anatomy looks weird!" "The faces are uncanny!" Hey so, you know the part where LLM generated images were tra…"But it's too detailed!" "The anatomy looks weird!" "The faces are uncanny!" Hey so, you know the part where LLM generated images wer…
↗Not briefed
hn
hn-ai
2moThe expensive model was cheaper: six other things I got wrong on an LLM judge—
↗Not briefed
hn
hn-ai
2moAsk HN: Anyone still coding like 2021? Where do you work?I can close an eye on using LLMs for research only.<p>These days I feel I am the only one that hasn’t changed their way of writing sof…
↗Not briefed
bluesky
bluesky-ai
2moUn ingeniero de Anthropic dimite y denuncia que las empresas de IA ocultan su riesgo real: "Están apostando nuestras vidas" www.eldiario.…Un ingeniero de Anthropic dimite y denuncia que las empresas de IA ocultan su riesgo real: "Están apostando nuestras vidas" www.eldi…
↗Not briefed
rss
the-decoder
2moDeepmind's AlphaGenome Atlas maps every possible DNA change in the human genomeGoogle Deepmind has used the AlphaGenome Atlas to predict what each of the roughly nine billion possible single-letter changes in the…
↗Not briefed
hn
hn-ai
2moWhat will our economic future look like?—
↗Not briefed
bluesky
bluesky-ai
2moHollywood CEO Security Costs Soar As Threats RiseHollywood CEO Security Costs Soar As Threats Rise
↗Not briefed
bluesky
bluesky-ai
2moLa advertencia de un ingeniero de IA de Antrophic que acaba de dimitir : "La IA podría matarnos a todos al final de esta década"La advertencia de un ingeniero de IA de Antrophic que acaba de dimitir : "La IA podría matarnos a todos al final de esta década"
↗Not briefed
bluesky
bluesky-ai
2mo…So Anthropic leadership not only misunderstood the TV show Person Of Interest *as a whole*, but they actively thought Team Samaritan were ……So Anthropic leadership not only misunderstood the TV show Person Of Interest *as a whole*, but they actively thought Team Samaritan…
↗Not briefed
arxiv
arxiv-ai
2moCoGe-GCD: Reframing Generalized Category Discovery with Compositional GeneralizationGeneralized Category Discovery (GCD) assigns unlabeled instances, mixed with labeled data, to known or novel categories, requiring hum…
↗Not briefed
hn
hn-ai
2moGambling with our lives': Another AI employee quits over safety concerns—
↗Not briefed
bluesky
bluesky-ai
2moAnthropic’s Economics team is sharing a new model of how AI might affect economic growth, jobs, wages, and more by 2030. Explore the scena…Anthropic’s Economics team is sharing a new model of how AI might affect economic growth, jobs, wages, and more by 2030. Explore the…
↗Not briefed
bluesky
bluesky-ai
2moWe should talk more about what a travesty the bibliography of OpenAI is 16 measly citations for a 165-page paper, of which 5 are just the …We should talk more about what a travesty the bibliography of OpenAI is 16 measly citations for a 165-page paper, of which 5 are just…
↗Not briefed
arxiv
arxiv-ai
2moFrom Retrieval to Weights: Parametric Individualization of Small Language Models with Individual Text CorporaWe approach a cognitive simulation perspective on episodic and semantic memory in multiple-choice question answering by incorporating…
↗Not briefed
arxiv
arxiv-ai
2moCompassOPD: Cross-Family On-Policy Distillation via Within-Family Likelihood ShiftsOn-policy distillation (OPD) provides dense token-level supervision on student-generated trajectories. Although OPD performs strongly…
↗Not briefed