frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Munich Laser Emits a Beam of Matter Waves (1999)

https://www.sciencedaily.com/releases/1999/03/990317060831.htm
1•peter_d_sherman•1m ago•0 comments

Anthropic CVP (Mythos) Defense Access Requirements Reduced to KYC

https://portal.anthropic.com/login?returnTo=%2Fprograms%2Fcvp#_
1•jcomp•1m ago•1 comments

New study narrows the timeline for the emergence of life on Earth

https://gizmodo.com/early-earth-was-a-molten-asteroid-blasted-hellscape-scientists-now-know-when-...
1•gmays•3m ago•0 comments

How Jensen Huang Built the Backdoor Route That Keeps Beijing in the AI Race

https://www.wsj.com/tech/jensen-huang-backdoor-ai-data-centers-a5e4499d
2•ViktorRay•8m ago•0 comments

Kiri – The Modern Container Runtime

https://runkiri.com/
3•nikolay•10m ago•0 comments

Show HN: Gawain – Require type hints on all Python variables

https://github.com/matheuzgomes/gawain
4•MarinhoD•11m ago•0 comments

Agent Substrate: The Core System

https://github.com/agent-substrate/substrate
3•thunderbong•13m ago•0 comments

Anthropic, can you please make up your mind with your models?

4•xlayn•13m ago•0 comments

Europe's Digital Wallets May Be More Traceable Than Intended

https://spectrum.ieee.org/europe-digital-id-security-cryptography
1•geox•14m ago•0 comments

Songs That Explain America

https://www.theatlantic.com/culture/2026/09/50-songs-explain-america-playlist/687598/
1•gmays•14m ago•0 comments

Training Your Replacement: Worker Incentives and the Pace of Automation

https://www.nber.org/papers/w35950
2•rohanm93•20m ago•0 comments

Discovery Feed for Agents

https://twitter.com/cadenbuild/status/2121372002313503003
1•cadenli2025•21m ago•0 comments

Smart AI Keyboard: Translate

https://play.google.com/store/apps/details?id=smart.ai.keyboard&hl=en_US
1•aikeyboard•22m ago•0 comments

FAA Outage: Why Duffy Is Spelled with Two Ms

https://www.flyingpenguin.com/faa-outage-why-duffy-is-spelled-with-two-ms/
1•zzzeek•29m ago•0 comments

Heading over Humanity

https://www.writingruxandrabio.com/p/handing-over-humanity
1•hwacth•34m ago•0 comments

Ranking 162 anime by how much of their dialogue is common words

https://etoku.app/anime/easiest
2•willmorrison•37m ago•3 comments

Show HN: Turbo-Browser-MCP – Batch forms, DOM deltas and submit guards for MCP

https://github.com/Tthanh2k6/turbo-browser-mcp
3•tthanh2k6•41m ago•0 comments

SRT Link to Red5 Opens Path to Real-Time Streaming Everywhere

https://www.red5.net/blog/srt-realtime-streaming-everywhere/
2•mondainx•42m ago•0 comments

Show HN: A blog on how computers work from scratch (1)

https://dinank.dev/posts/between-wall-and-first-fetch.html
2•daynum•43m ago•2 comments

A Plan of the Cities of London and Westminster, and Borough of Southwark

https://en.wikipedia.org/wiki/John_Rocque%27s_maps_of_London
1•arbuge•43m ago•0 comments

Musings on long-context fine-tuning

https://hardikmeisheri.github.io/blog/musings-long-context-finetuning.html
1•eigenBasis•45m ago•0 comments

I hope you die: Inside Jensen Huang's reported tirades at Nvidia board meetings

https://www.indiatoday.in/world/story/jensen-huang-nvidia-workplace-allegations-executive-outburs...
3•gsky•46m ago•0 comments

Flattening Every Memory Peak in Long-Context Mixture-of-Experts Training

https://arxiv.org/abs/2609.14306
2•gmays•48m ago•0 comments

Show HN: KrocPDF – A Local-First, Alternative to iLovePDF

https://github.com/KrocPDF/KrocPDF
2•abhijeetcreates•49m ago•0 comments

Anthropic bought a one-person, AI-powered biotech

https://www.nature.com/articles/d41586-026-03640-9
1•nate•49m ago•0 comments

A/B test whether pre-gathered context helps coding agents complete tasks

https://github.com/unblocked/context-engine-simulator
1•fagnerbrack•49m ago•0 comments

India applying inhumane tech illegally on Canadian civilians on Canadian soil

2•whiteangel1984•49m ago•0 comments

SuperMimic: Compute is no longer the bottleneck to robotics

https://supermimic.github.io/
2•maccaw•51m ago•0 comments

What Math Means to Me

https://arxiv.org/abs/2609.32028
1•fanyangxyz33•53m ago•0 comments

Linus Torvalds Wonders If Linux's AI-Prompted Patch Flood Has Peaked

https://linux.slashdot.org/story/26/11/14/0827227/linus-torvalds-wonders-if-linuxs-ai-prompted-pa...
2•mfiguiere•54m ago•0 comments
Open in hackernews

Anthropic, can you please make up your mind with your models?

2•xlayn•13m ago
Anthropic, can you please make up your mind with the models?

I don't know if this is related to the IPO, fable 5.2 is WAY more expensive to run, if you put it in xhigh you can use in an 1 or 2 hours the allocation of the whole week... so you gave "free" resets for fable... to make "easy" on people that the subscription is going to start lasting less and less...

So I went back to fable 5.1 and the dumbification has started to happen AGAIN... when I want the nerd that writes 10 pages of DENSE text non stop I go to opus, fable is more expensive because it has better presets, but now fable 5.1 is starting to go into the uber nerd in adderal area too PLUS starting to do dumb things... I'm in this process of running something, get logs, tune, get logs, tune... and I told the thing multiple tmes... hey create scripts if you need to run something multiple times... the thing burned like crazy in fable 5.1 effort medium... until you put in caps in expletives what's going on... then writes the script and 16 tokens to check the results...

I repent that I forgot to cancel the friking thing last month... PLEASE STOP THE A/B THING... I don't pay you 100$ 80$ 65$ 70$... I pay always the same 100$ and I want the 100$ service from you each time...

venting here because all the ai companies support is non existent