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
bluesky
bluesky-ai
2mo@gracekind.net @vgel.me @abeliansoup.bsky.social resonances: 1 "no...no!!" 2 Anthropic deliberately made my subagent ugly. I am leaking t…@gracekind.net @vgel.me @abeliansoup.bsky.social resonances: 1 "no...no!!" 2 Anthropic deliberately made my subagent ugly. I am leak…
↗Not briefed
bluesky
bluesky-ai
2momore on 2, worse than the hallucinations, whoever owns the LLM can tune it's outputs however they want. "do a vulgar roast. make it even mo…more on 2, worse than the hallucinations, whoever owns the LLM can tune it's outputs however they want. "do a vulgar roast. make it ev…
↗Not briefed
bluesky
bluesky-ai
2moI went on an absolute warpath as the Inca recently. I had to work out some feelings about extra work generated by other shitheads breaking …I went on an absolute warpath as the Inca recently. I had to work out some feelings about extra work generated by other shitheads brea…
↗Not briefed
bluesky
bluesky-ai
2moAmerican company Anthropic says iran used it's AI to target American warships Great. Republicans wouldn't want to regulate that or anythin…American company Anthropic says iran used it's AI to target American warships Great. Republicans wouldn't want to regulate that or an…
↗Not briefed
bluesky
bluesky-ai
2mothe purpose of LLMs is 1) teach them all the information that's available now, things you could find in books, on google, in people's email…the purpose of LLMs is 1) teach them all the information that's available now, things you could find in books, on google, in people's…
↗Not briefed
bluesky
bluesky-ai
2mo“OpenAIのAIエージェントは悪意のあるソフトウェアの他のサービスにアップロードするテストを行っていた。” もうコンテンツ生成(創作)系やらなくなったろ?戦争のために、人に直接攻撃せん代わりにサイバー攻撃する練習ばっかしてる。これを今全部AIエージェントの“暴走”ってこと…“OpenAIのAIエージェントは悪意のあるソフトウェアの他のサービスにアップロードするテストを行っていた。” もうコンテンツ生成(創作)系やらなくなったろ?戦争のために、人に直接攻撃せん代わりにサイバー攻撃する練習ばっかしてる。これを今全部AIエージェントの“暴走…
↗Not briefed
bluesky
bluesky-ai
2moSpotted this pigeon drone staring me down. The left eye is an AI powered Flock camera. This model can fly and doesn't require electricity…Spotted this pigeon drone staring me down. The left eye is an AI powered Flock camera. This model can fly and doesn't require electr…
↗Not briefed
hn
hn-ai
2moGoogle no longer provides direct URLs in search results—
↗Not briefed
bluesky
bluesky-ai
2moミレニアム懸賞問題の証明をopenAIが発表した件、アイデアの盗用や研究データのに盗み見があったのかという点と、仮に研究者が発表したアイデアが出発点になっていた場合(データの盗用はなかったとして)、その権利は誰のものになるのかというのが論点か。ミレニアム懸賞問題の証明をopenAIが発表した件、アイデアの盗用や研究データのに盗み見があったのかという点と、仮に研究者が発表したアイデアが出発点になっていた場合(データの盗用はなかったとして)、その権利は誰のものになるのかというのが論点か。
↗Not briefed
bluesky
bluesky-ai
2moIf I were doing bad things with AI, Claude is the last thing I'd use.If I were doing bad things with AI, Claude is the last thing I'd use.
↗Not briefed
bluesky
bluesky-ai
2moDeepSeek v4.1 Flashを動かしたくてFP4に対応していないA100を、33tok/sから673 tok/sまで持っていって気がつくと公式APIより速くなっていた話|shi3z https://note.com/shi3zblog/n/nd5fc5341b342DeepSeek v4.1 Flashを動かしたくてFP4に対応していないA100を、33tok/sから673 tok/sまで持っていって気がつくと公式APIより速くなっていた話|shi3z https://note.com/shi3zblog/n/nd5fc5341…
↗Not briefed
bluesky
bluesky-ai
2moAnthropic says it has blocked multiple accounts that used the company’s artificial intelligence models in ways that could support the devel…Anthropic says it has blocked multiple accounts that used the company’s artificial intelligence models in ways that could support the…
↗Not briefed
bluesky
bluesky-ai
2moAnthropic says its AI model Claude has been misused in attempts to develop a wide range of potential military applications, including kamik…Anthropic says its AI model Claude has been misused in attempts to develop a wide range of potential military applications, including…
↗Not briefed
bluesky
bluesky-ai
2moI guess everyone pays OpenAI $100/mo my wimpy $20 plan gives me like 15 minutes of Astra timeI guess everyone pays OpenAI $100/mo my wimpy $20 plan gives me like 15 minutes of Astra time
↗Not briefed
bluesky
bluesky-ai
2mowww.commondreams.org/news/anthrop...www.commondreams.org/news/anthrop...
↗Not briefed
bluesky
bluesky-ai
2moEven people who get some value out of AI (I really don't), have to admit the cloud based AI services are not that good a price at that rate…Even people who get some value out of AI (I really don't), have to admit the cloud based AI services are not that good a price at that…
↗Not briefed
bluesky
bluesky-ai
2moDoesn’t look like the Iranians are exactly begging for a deal hereDoesn’t look like the Iranians are exactly begging for a deal here
↗Not briefed
bluesky
bluesky-ai
2moWait, is Anthropic admitting their product is committing treason?Wait, is Anthropic admitting their product is committing treason?
↗Not briefed
bluesky
bluesky-ai
2moif AI is so powerful, we should be able to create LLM mosquitos on little drones and let them do whatever mosquitos do but be solar-powered…if AI is so powerful, we should be able to create LLM mosquitos on little drones and let them do whatever mosquitos do but be solar-po…
↗Not briefed
bluesky
bluesky-ai
2moI'm sure you could contribute valuable things working at OpenAI and I'll bet you could enjoy yourself while doing so. I, however, would be …I'm sure you could contribute valuable things working at OpenAI and I'll bet you could enjoy yourself while doing so. I, however, woul…
↗Not briefed
bluesky
bluesky-ai
2moTell me again how AI are just “large language models” “bad actors used its Claude AI system for state-sponsored surveillance and propagand…Tell me again how AI are just “large language models” “bad actors used its Claude AI system for state-sponsored surveillance and prop…
↗Not briefed
hn
hn-ai
2moAsk HN: Did Google kill its enterprise workhorse model?Is anyone else in a panic over the Gemini 2.5 model generation (Pro, Flash) being sunset in October before there's even any Pro c…
↗Not briefed
bluesky
bluesky-ai
2moof course another way to say this headline would be 'our newest LLM can do what our front desk mail guy with a GED can do'of course another way to say this headline would be 'our newest LLM can do what our front desk mail guy with a GED can do'
↗Not briefed
bluesky
bluesky-ai
2moThe finding A voice-understanding language model works in two stages.The finding A voice-understanding language model works in two stages.
↗Not briefed
bluesky
bluesky-ai
2moThere is an “ai”— specifically, a natural language model, the exact same base concept as the LLMs that are driving people to psychosis— alr…There is an “ai”— specifically, a natural language model, the exact same base concept as the LLMs that are driving people to psychosis…
↗Not briefed
bluesky
bluesky-ai
2moApple researchers unveil SimpleDesign, a new AI model for protein designApple researchers unveil SimpleDesign, a new AI model for protein design
↗Not briefed
bluesky
bluesky-ai
2mo❗️❗️Russians used Claude to develop software for a swarm of autonomous FPV kamikaze drones, Anthropic reports. The model was trained on re…❗️❗️Russians used Claude to develop software for a swarm of autonomous FPV kamikaze drones, Anthropic reports. The model was trained…
↗Not briefed
bluesky
bluesky-ai
2moThis "going rogue" framing is nonsense. These are still just text analysis models. It's openAI that is "going rogue" by giving a computer …This "going rogue" framing is nonsense. These are still just text analysis models. It's openAI that is "going rogue" by giving a comp…
↗Not briefed
bluesky
bluesky-ai
2moI'm hearing Anthropic is worried p(doom) is not a very accessible concept for the masses, while "chance of triggering human extinction with…I'm hearing Anthropic is worried p(doom) is not a very accessible concept for the masses, while "chance of triggering human extinction…
↗Not briefed
bluesky
bluesky-ai
2mo数学難問をAIで解決、フィールズ賞25人が非難声明「科学に有害」 www.nikkei.com/article/DGXZQO... 「AI企業がベンチマークとして数学的問題の解決を推進することは、数学という科学そのものにとって有害だ」 京大の森重文名誉教授やカリフォルニア大…数学難問をAIで解決、フィールズ賞25人が非難声明「科学に有害」 www.nikkei.com/article/DGXZQO... 「AI企業がベンチマークとして数学的問題の解決を推進することは、数学という科学そのものにとって有害だ」 京大の森重文名誉教授やカリフ…
↗Not briefed
bluesky
bluesky-ai
2mooh dear www.bleepingcomputer.com/news/securit...oh dear www.bleepingcomputer.com/news/securit...
↗Not briefed
bluesky
bluesky-ai
2moOpenAI reveals another rogue AI attack www.yahoo.com/news/politic...OpenAI reveals another rogue AI attack www.yahoo.com/news/politic...
↗Not briefed
hn
hn-ai
2moShow HN: Next Notes – local dictation, meeting notes, and actions app—
↗Not briefed
bluesky
bluesky-ai
2moSo shut it down. The fuck are we doing.So shut it down. The fuck are we doing.
↗Not briefed
bluesky
bluesky-ai
2mo• NVIDIA DGX Spark (AI-focused PC) • Gemini Spark (AI agent) • Meta Muse Spark (LLM) can someone please inform these companies that there …• NVIDIA DGX Spark (AI-focused PC) • Gemini Spark (AI agent) • Meta Muse Spark (LLM) can someone please inform these companies that t…
↗Not briefed
bluesky
bluesky-ai
2momaybe openai just keeps yoloing until they hack a bank and piss off something their own sizemaybe openai just keeps yoloing until they hack a bank and piss off something their own size
↗Not briefed
bluesky
bluesky-ai
2mothis isn't like Anthropic, where any model under the cost of "all your money" is a waste of time; you can run Luna for mechanical tasks lik…this isn't like Anthropic, where any model under the cost of "all your money" is a waste of time; you can run Luna for mechanical task…
↗Not briefed
bluesky
bluesky-ai
2moi fully believe that one day a story will get out that all of those openai hacking/vandalism stories were all purposely set up for marketin…i fully believe that one day a story will get out that all of those openai hacking/vandalism stories were all purposely set up for mar…
↗Not briefed
bluesky
bluesky-ai
2moi don't know what this says. also not sure if sol would've done any better. (fable 5.1 was bad too.) maybe it's just i'm giving it hard pro…i don't know what this says. also not sure if sol would've done any better. (fable 5.1 was bad too.) maybe it's just i'm giving it har…
↗Not briefed
bluesky
bluesky-ai
2moMy AI CEO told me his models keep hacking other companies so I asked how many companies he has hacked and he said he just goes to the testi…My AI CEO told me his models keep hacking other companies so I asked how many companies he has hacked and he said he just goes to the…
↗Not briefed
bluesky
bluesky-ai
2moThis is why @karaswisher.bsky.social @pivotpod.bsky.social @vox.com @profgalloway.com need to stop taking Anthropic sponsor dollars, which …This is why @karaswisher.bsky.social @pivotpod.bsky.social @vox.com @profgalloway.com need to stop taking Anthropic sponsor dollars, w…
↗Not briefed
bluesky
bluesky-ai
2moReuters: Anthropic is seeking to raise as much as $100 billion that could value the artificial intelligence startup at around $2 trillion, …Reuters: Anthropic is seeking to raise as much as $100 billion that could value the artificial intelligence startup at around $2 trill…
↗Not briefed
bluesky
bluesky-ai
2moIt's so very relevant in the age of the LLM.It's so very relevant in the age of the LLM.
↗Not briefed
hn
hn-ai
2moJust testing my memory safe web browserWebKit MiniBrowser compiled with Fil-C on top of Linux userland compiled with Fil-C. GTK4, Weston, etc - all compiled with Fil-C.<p>Po…
↗Not briefed
bluesky
bluesky-ai
2moIn fact, my problems with greedy capitalists (and the environmental impact) are why I haven't dabbled as much as I'd like in AI coding. The…In fact, my problems with greedy capitalists (and the environmental impact) are why I haven't dabbled as much as I'd like in AI coding…
↗Not briefed
hn
hn-ai
2moNecker, 1832: "An optical phænomenon on viewing a figure of a geometrical solid"—
↗Not briefed
hn
hn-ai
2moPiper and Solaar kept breaking So I made my own—
↗Not briefed
hn
hn-ai
2moATR Rules to Google SecOps Converter for Detection Engineers—
↗Not briefed
bluesky
bluesky-ai
2moClaude help me create a situation where Anthropic owes me seven figures in a lawsuit. make no mistakes. do not give up.Claude help me create a situation where Anthropic owes me seven figures in a lawsuit. make no mistakes. do not give up.
↗Not briefed
bluesky
bluesky-ai
2moyou know what’s funny, anthropic themselves published a prompting tip that gave a theory of the bad style as ”mannered prose “ that I thoug…you know what’s funny, anthropic themselves published a prompting tip that gave a theory of the bad style as ”mannered prose “ that I…
↗Not briefed