Bento
HN Top
Top Hacker News stories for AI builders, hackers, and founders.
Judge rules DOGE cancellation of humanities grants was unconstitutional
If Anyone Builds It, Everyone Dies
Compound drivers of Antarctic sea ice loss and Southern Ocean destratification
Teaching Claude Why
Non-determinism is an issue with patching CVEs
Looking at the data behind prediction markets
Show HN: GETadb.com – every GET request creates a DB
What we lost the last time code got cheap
Hallucinations Undermine Trust; Metacognition Is a Way Forward
Show HN: Git for AI Agents
Tesla is recalling its cheaper Cybertruck because the wheels might fall off
Poland is now among the 20 largest economies
GPT-5.5 Price Increase: What It Costs
A polynomial autoencoder beats PCA on transformer embeddings
Mojo 1.0 Beta
Agentic Engineering
How to make SSE token streams resumable, cancellable, and multi-device
Hardening Firefox with Claude Mythos Preview
The Traveling Salesdog Problem
Nonprofit hospitals spend billions on consultants with no clear effect
Researchers discover advanced language processing in the unconscious human brain
Two Home Affairs officials suspended after AI 'hallucinations' found in policy paper
Mozilla says 271 vulnerabilities found by Mythos and "almost no false positives"
Show HN: Kstack – Skill pack for monitoring/troubleshooting K8s in Claude Code
AI Slop is Killing Online Communities
Natural Language Autoencoders: Turning Claude's Thoughts into Text
Show HN: Stage CLI – a tool to make reading your AI generated changes easier
DeepSeek 4 Flash local inference engine for Metal
Only 173? Tesla Cybertruck RWD Recall Reveals Low Sales
Why should a Trace-ID be 128 bits? (A Surprisingly Long Answer)
AlphaEvolve: Gemini-powered coding agent scaling impact across fields
37x Speedup in Lattice Boltzmann Cylinder Flow
Agent-harness-kit scaffolding for multi-agent workflows (MCP, provider-agnostic)
ZAYA1-8B Matches DeepSeek-R1 on Math with Less Than 1B Active Parameters
How Unsloth and Nvidia made LLM training 25% faster on consumer GPUs
Show HN: Agent-skills-eval – Test whether Agent Skills improve outputs
ProgramBench: Can Language Models Rebuild Programs from Scratch?
The Mathematical Dance Inside Plant Cells
What British people mean when they say 'sorry'
Finding the differences in a series of power supplies
Perturb-MARS: Reading mouse experiments through a human lens
Programming Still Sucks
Following the Text Gradient at Scale
David Sacks crashed and burned in the White House
Mythos is the best cybersecurity news in a decade
MIT's virtual violin offers luthiers a new design tool
Learning the Integral of a Diffusion Model
Proton Meet
