frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

Open in hackernews

AI "reasoning" models don't reason at all

https://twitter.com/RubenHssd/status/1931389580105925115
3•doener•4h ago

Comments

aurizon•4h ago
AI 'reasoning' is a allied to password cracking, where every combination of the allowed symbol set is incremented across until the locked door is opened. Analogously the AI does a similar task with the ability to create and test code until one works, and similarly hack code with a blizzard of variances until an entry point(bug) is found = record and continue. The speed and parallelism is not a reasoned path, it is an exhaustive path that, hopefully, finds the hole. It also finds 'hallucinations' - which a reasoning mind recognises, but these early stage AI's often do not.
JPLeRouzic•3h ago
LLMs are not based on brute force algorithms. They are based on finding the best string of tokens (for text based LLMs) that would correspond to a prompt.

To illustrate that you can see how DeepSeek "thinks". There are lots of "Given the instructions, if we", "Alternatively, we can ", "Also, if ", "wait", "but" etc...

Show HN: I vibe coded an app for my 2 years old niece using Claude Code

https://www.youtube.com/watch?v=JRkprvGAK0Y
1•EtishaGarg•5m ago•0 comments

AdLinkr – See Every Competitor LinkedIn Ad in One Place with Insights

1•praveen_ravi•6m ago•0 comments

Activity annealing leads to a ductile-to-brittle transition in amorphous solids

https://www.nature.com/articles/s41567-024-02724-5
2•PaulHoule•9m ago•0 comments

Tapping into the World’s largest gold reserves

https://www.uni-goettingen.de/en/3240.html?id=7808
1•geox•10m ago•0 comments

I'm pivoting because LLMs aren't ready for safe clinical use by patients

https://amirkiani.xyz/posts/yari-care-nav/
2•amiroo•12m ago•0 comments

Show HN: I built FoundersWall where builders pin chaos instead of chasing likes

https://founderswall.com
1•harvansh•14m ago•1 comments

When the Assistant Becomes the Attacker: Hidden Risks of Tool-Enabled LLMs

https://www.sakana.fr/blog/2025-06-08-llm-hidden-risks/
1•killermouse0•15m ago•0 comments

Why Philosophers Should Care About Computational Complexity [video]

https://www.youtube.com/watch?v=OST1DjD08Hg
1•amichail•20m ago•0 comments

The Problem with Shadow Development

https://leaddev.com/technical-direction/the-problem-with-shadow-development
1•kiyanwang•23m ago•0 comments

IPWT: A unified, computationally feasible framework for consciousness

https://github.com/dmf-archive/IPWT
2•NetRunnerSu•23m ago•0 comments

How I Use AI (2025)

https://blog.sshh.io/p/how-i-use-ai-2025
1•sshh12•26m ago•0 comments

US budget cuts could derail global science collaborations

https://www.nature.com/articles/d41586-025-01764-y
2•rntn•27m ago•1 comments

MariaDB 11.8 LTS Released

https://mariadb.org/11-8-lts-released/
4•gslin•28m ago•0 comments

The Dogs of Chernobyl Are Experiencing Rapid Evolution, Study Suggests

https://www.popularmechanics.com/science/animals/a64969223/chernobyl-dogs-dna-rapid-evolution/
4•Ozarkian•29m ago•1 comments

Building Unreal Engine 5.6 from the GitHub Source Code on GNU/Linux

https://www.youtube.com/watch?v=eGsg3Ph3QPA
1•NuLL3rr0r•31m ago•0 comments

Show HN: FansFinder – A Search Engine for OnlyFans Creators

1•FnasFinder•31m ago•0 comments

Big Bets

https://jackdanger.com/big-bets/
1•kiyanwang•32m ago•0 comments

The USB-C Dream is Dead

https://www.androidauthority.com/usb-c-cant-be-fixed-3560127/
3•hannofcart•32m ago•0 comments

How I Program with Agents

https://sketch.dev/blog/programming-with-agents
3•sqs•33m ago•0 comments

Rkyv (archive) is a zero-copy deserialization framework for Rust

https://rkyv.org/rkyv.html
2•ChadNauseam•36m ago•0 comments

Microsoft Needs a SteamOS Competitor, and It Needs It Yesterday

https://www.howtogeek.com/microsoft-needs-a-steamos-competitor-and-it-needs-it-yesterday/
1•dingi•36m ago•0 comments

Binghamton researchers make dissolvable battery using probiotics

https://www.binghamton.edu/news/story/5482/binghamton-university-researchers-make-dissolvable-battery-using-probiotics
1•giuliomagnifico•38m ago•1 comments

Your Smartphone Is a Parasite, According to Evolution

https://www.nakedcapitalism.com/2025/06/your-smartphone-is-a-parasite-according-to-evolution.html
6•cratermoon•40m ago•0 comments

Comma v0.1 1T and 2T–7B LLMs trained on openly licensed text

https://simonwillison.net/2025/Jun/7/comma/
1•amrrs•40m ago•0 comments

How to run LLMs locally on mobile devices (with Gemma and On-Device AI tools)

https://annjose.com/post/mobile-on-device-ai-hands-on-gemma/
1•annjose•42m ago•1 comments

Meme Stocks Made Him a Fortune. Now He's Betting on Flying Taxis

https://www.wsj.com/business/meme-stocks-made-him-a-fortune-now-hes-betting-on-flying-taxis-63449215
1•Bostonian•43m ago•1 comments

AI-Generated Code Statistics 2025: Is Your Developer Job Safe?

https://www.elitebrains.com/blog/aI-generated-code-statistics-2025
1•mathattack•43m ago•1 comments

Tech giants in the Gulf are building super-apps to dominate digital life

https://restofworld.org/2025/middle-east-super-apps/
2•thm•43m ago•0 comments

Ask HN: How long do you ever leave your server running without updates?

1•bagol•43m ago•0 comments

How to Annotate Everything

https://beepb00p.xyz/annotating.html
1•Tomte•44m ago•0 comments