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
1dMost of these shiny LLM vuln reports are produced by either 1) the LLM companies, 2) in-house researchers at cybersecurity industry compani…Most of these shiny LLM vuln reports are produced by either 1) the LLM companies, 2) in-house researchers at cybersecurity industry co…
↗Not briefed
hn
hn-ai
1dI Moved from Debian to NixOS—
↗Not briefed
bluesky
bluesky-ai
1d“Just imagine if Jimi Hendrix had been asked to post-edit an #AI cover of Bob Dylan's ‘All Along the Watchtower’ instead of creating his un…“Just imagine if Jimi Hendrix had been asked to post-edit an #AI cover of Bob Dylan's ‘All Along the Watchtower’ instead of creating h…
↗Not briefed
hn
hn-ai
1dAnthropic IPO at $83—
↗Not briefed
hn
hn-ai
1dShow HN: My 1990 cellular automata HyperCard stack runs again, XCMDs in WASM—
↗Not briefed
arxiv
arxiv-ai
1dLearn, Freeze, and Transfer an Activation Function Across ArchitecturesTransfer learning typically reuses knowledge acquired by a source network through its learned weights. We investigate whether a learne…
↗Not briefed
arxiv
arxiv-ai
1dLearning to Think with Images Faithfully via Adaptive Data SelectionThinking with images (TwI) empowers vision-language models (VLMs) to overcome perceptual bottlenecks by actively cropping and zooming…
↗Not briefed
hn
hn-ai
1dShow HN: Chipload – Open-source 2.5D CAM for CNC routingThis is an Open source 2.5D CAM software I made (w. Claude) for generating gcode for CNC routers. It is quite full-featured by now and…
↗Not briefed
arxiv
arxiv-ai
1dBackbone Matters: Benchmarking Encoder-Decoder Networks for Semantic Segmentation of Driving Scenes on BDD100KSemantic segmentation gives an autonomous vehicle a dense, pixel-level description of its surroundings and is a core part of scene und…
↗Not briefed
bluesky
bluesky-ai
1dI’m seeing fetishisation of “African” writing as if it is this weird magical thing that is at the same time (patronisingly) always wonderfu…I’m seeing fetishisation of “African” writing as if it is this weird magical thing that is at the same time (patronisingly) always won…
↗Not briefed
arxiv
arxiv-ai
1dInvertible Transport Between Sample-Only Distributions via Projected Characteristic FunctionsMany useful latent distributions can be sampled but do not have a tractable density. Such distributions may be represented by empirica…
↗Not briefed
bluesky
bluesky-ai
1dHeard this earlier this morning, pretty interesting. *ANTHROPIC IPO SAID TO GIVE NOTHING TO MOST INSTITUTIONAL BUYERSHeard this earlier this morning, pretty interesting. *ANTHROPIC IPO SAID TO GIVE NOTHING TO MOST INSTITUTIONAL BUYERS
↗Not briefed
arxiv
arxiv-ai
1dReward misalignment as an opportunity for materials discoveryReward misalignment is usually regarded as a failure of artificial intelligence. In materials discovery with generative AI models, how…
↗Not briefed
arxiv
arxiv-ai
1dThermodynamically Consistent Neural Networks for a Universal Fluid Gibbs Free Energy Equation of StateOver the past three centuries, many thermodynamics researchers have worked to develop fluid equations of state with both high accuracy…
↗Not briefed
rss
techmeme
1dThe European Commission confirms a cyberattack that disrupted systems for 10+ days; CERT-EU says international hackers used AI penetration testing tool ARTEX (Gerardo Fortuna/Politico)Gerardo Fortuna / Politico: The European Commission confirms a cyberattack that disrupted systems for 10+ days; CERT-EU says internati…
↗Not briefed
bluesky
bluesky-ai
1dCan I say I find gendering an LLM strange. Like saying he instead of it or they.Can I say I find gendering an LLM strange. Like saying he instead of it or they.
↗Not briefed
bluesky
bluesky-ai
1dAnthropic has absolutely pushed the idea that they're the good and thoughtful AI company, but anyone who's gullible enough to have fallen f…Anthropic has absolutely pushed the idea that they're the good and thoughtful AI company, but anyone who's gullible enough to have fal…
↗Not briefed
hn
hn-ai
1dShow HN: ChartKit – a chart library where the agent writes JSON, not code—
↗Not briefed
hn
hn-ai
1dAptera Delivers First Vehicle to Customer—
↗Not briefed
hn
hn-ai
1dShow HN: Wiretuple – Cut LLM structured-output tokens with positional tuples—
↗Not briefed
bluesky
bluesky-ai
1d🔴 URGENT - Anthropic entre en Bourse : la plus grande levée de fonds de l’Histoire commence🔴 URGENT - Anthropic entre en Bourse : la plus grande levée de fonds de l’Histoire commence
↗Not briefed
bluesky
bluesky-ai
1dAnyone who thinks Anthropic has a good guy reputation probably shouldn't be making their own financial decisions.Anyone who thinks Anthropic has a good guy reputation probably shouldn't be making their own financial decisions.
↗Not briefed
bluesky
bluesky-ai
1di can't explain how fucking tired i am of talking about AI. i took 10 uber rides last week and do you know how many of them turned to AI?? …i can't explain how fucking tired i am of talking about AI. i took 10 uber rides last week and do you know how many of them turned to…
↗Not briefed
bluesky
bluesky-ai
1dTimes Square.Times Square.
↗Not briefed
bluesky
bluesky-ai
1dthis is the kind of place where the federal gov puts its (non-LLM) data centers by the way. we've had one here since the 2010s. nobody feel…this is the kind of place where the federal gov puts its (non-LLM) data centers by the way. we've had one here since the 2010s. nobody…
↗Not briefed
bluesky
bluesky-ai
1d確かにかつて(ほとんどの)プログラマーは機械語を書くことをコンパイラに任せて手放したし、メモリを管理することも高級言語のランタイムに任せて手放してきたけど、「ビジネスロジックの実装をLLMに任せて手早す」のは、コンパイラやランタイムほどには出力が信用できないというところが厄介な…確かにかつて(ほとんどの)プログラマーは機械語を書くことをコンパイラに任せて手放したし、メモリを管理することも高級言語のランタイムに任せて手放してきたけど、「ビジネスロジックの実装をLLMに任せて手早す」のは、コンパイラやランタイムほどには出力が信用できないというとこ…
↗Not briefed
arxiv
arxiv-ai
1dRegionAgent: Two-Stage Preference Optimization for Regional Image RestorationImages under non-uniform coupled degradations require diverse restoration capabilities for different degraded regions. Existing method…
↗Not briefed
hn
hn-ai
1dShow HN: Exploitwatch, get alerts for the CVEs in your containers worth fixing—
↗Not briefed
hn
hn-ai
1dAnthropic IPO: Live—
↗Not briefed
hn
hn-ai
1dRant: Gemini 4 does not make any senseOn November 18th, Google released another model, this one called Neon. Making it the fourth model in just 2 months.<p>OpenAI makes sen…
↗Not briefed
hn
hn-ai
1dLet decisions API control NetHack monsters—
↗Not briefed
hn
hn-ai
1dSuspected Chinese APT exploited VMware vCenter zero-day to bypass auth—
↗Not briefed
hn
hn-ai
1dFastUnlock: Windows Narrator LockScreen Bypass 0day Vulnerability—
↗Not briefed
rss
theverge-ai
1dWith the biggest IPO of all time, Anthropic’s ‘good guy’ reputation is on the lineToday is a new beginning for Anthropic: a years-in-the-making IPO that beat out SpaceX for most valuable of all time, having raised $8…
↗Not briefed
bluesky
bluesky-ai
1dIt's not anti-AI leftists saying this. Microsoft Director of Applied Science Brent Hecht warned in internal documents that scraping news fo…It's not anti-AI leftists saying this. Microsoft Director of Applied Science Brent Hecht warned in internal documents that scraping ne…
↗Not briefed
hn
hn-ai
1dShow HN: Spec-receipt – does your code still match your spec?—
↗Not briefed
hn
hn-ai
1dShow HN: Agentest – offline tests for AI agents from OpenTelemetry traces—
↗Not briefed
hn
hn-ai
1dReverseRosetta—
↗Not briefed
bluesky
bluesky-ai
1dSeems like the availability AI tools is making it clearer that "for-profit journalism" is an untenable business model.Seems like the availability AI tools is making it clearer that "for-profit journalism" is an untenable business model.
↗Not briefed
bluesky
bluesky-ai
1dIt'll make money of course but it's more evidence that greater fool >>> economic value. Target 12% DCF and assume a generous 26% margin, an…It'll make money of course but it's more evidence that greater fool >>> economic value. Target 12% DCF and assume a generous 26% margi…
↗Not briefed
arxiv
arxiv-ai
1dBudget Internalization: Fixed-Budget RLVR Trades Test-Time Scaling for In-Budget AccuracyReasoning models often become more accurate when they can generate longer answers at test time. This behavior is known as test-time sc…
↗Not briefed
arxiv
arxiv-ai
1dThe FrontierOR Challenge: Agentic Algorithm Design for Operations ResearchOperations research (OR) relies on algorithms that experts tailor to each problem: a new routing, scheduling or network design problem…
↗Not briefed
bluesky
bluesky-ai
1dCurrent detection tools don’t take “white writing”’as the norm and then classify anything that deviates from that as LLM. That is, after …Current detection tools don’t take “white writing”’as the norm and then classify anything that deviates from that as LLM. That is, a…
↗Not briefed
bluesky
bluesky-ai
1dBrilliant A bash script that writes to Sam Altman once a day to report the RAM he missedBrilliant A bash script that writes to Sam Altman once a day to report the RAM he missed
↗Not briefed
hn
hn-ai
1dPHP-FPM request smuggling allows unauthenticated remote code execution—
↗Not briefed
rss
techmeme
1dA UK tribunal rules against the UK Home Office, saying it must acknowledge issuing a secret "technical capability notice" to Apple to bypass its encryption (Tim Bradshaw/Financial Times)Tim Bradshaw / Financial Times: A UK tribunal rules against the UK Home Office, saying it must acknowledge issuing a secret “technical…
↗Not briefed
arxiv
arxiv-ai
1dNariQ: Sub-3-Bit Quantization for Long-Horizon Agentic TasksLong-horizon agentic tasks require a model to call tools over many steps at contexts of 100K tokens or more, yet many low-bit quantiza…
↗Not briefed
bluesky
bluesky-ai
1dChrist, this is bleak.Christ, this is bleak.
↗Not briefed
bluesky
bluesky-ai
1dNEW ODD LOTS: The problem with AI’s most bullish chart. On this day of the Anthropic IPO, we talk about why a continued rise in GPU price…NEW ODD LOTS: The problem with AI’s most bullish chart. On this day of the Anthropic IPO, we talk about why a continued rise in GPU…
↗Not briefed
bluesky
bluesky-ai
1dAnthropic:Anthropic:
↗Not briefed