Pyyan / Trending / Hacker News

AI on Hacker News

See it on Hacker News

The AI stories the engineering audience actually upvoted. · ranked by Hacker News points · pulled 15d ago

6 Leading in each categoryThe top entry from every group, so you can scan the shape of it before the detail

30 General

E#3
US citizen charged after GrapheneOS phone wipes during airport searchHacker News
1340 points1120 comments33d ago
F#6
AI is removing the middle class of software engineering?Hacker News
956 points881 comments16d ago
A#7
As AI eats the web, the internet’s collective memory is disappearingHacker News
933 points969 comments18d ago
A#10
AI companies are shredding rare booksHacker News
801 points513 comments32d ago
M#11
AI-Generated Images Discourage Me from Reading Your BlogHacker News
796 points465 comments24d ago
E#14
Docker Sandboxes – Disposable, isolated sandboxes for AI agentsHacker News
690 points397 comments18d ago
R#17
Mark Zuckerberg attacks 'closed' AI rivals as Meta returns to open modelsHacker News

https://archive.is/20LOJ https://www.meta.com/thefutureisforeveryone/

639 points601 comments18d ago
Y#19
AI's top startups are barely publishing their researchHacker News
619 points318 comments30d ago
G#21
Google fixed more Chrome bugs in June than over the past two years, thanks to AIHacker News
585 points610 comments28d ago
D#25
Oracle bans AI-generated code from OpenJDKHacker News
535 points382 comments21d ago
M#30
How Claude marks AI-generated contentHacker News
445 points414 comments18d ago
0#33
Go is an ideal language for AI-assisted software engineeringHacker News
434 points526 comments17d ago
Y#35
Software development with AI is starting to feel like cooking steakHacker News
420 points421 comments22d ago
T#36
Open-weight AI is having its Kubernetes momentHacker News
416 points322 comments34d ago
C#39
Document-borne AI worms can self-propagate through Copilot for WordHacker News
384 points300 comments30d ago
M#40
The AI AestheticHacker News
378 points176 comments29d ago
A#42
GCC steering committee announces AI policyHacker News
357 points429 comments29d ago
F#43
AI financial advice is surprisingly good, especially if you ask right questionsHacker News
351 points403 comments27d ago
W#45
Humans missed 1 in 3 threats approving AI agent commands across 40k game runsHacker News
339 points245 comments22d ago
M#47
Meta Ran Ads That Contained AI-Generated Child Sexual Abuse ImageryHacker News
325 points270 comments23d ago
M#49
Managing AI Coding Costs at ScaleHacker News
315 points268 comments21d ago
P#52
What is happening to jobs? Separating AI hype from realityHacker News
302 points380 comments34d ago
G#54
Someone is running mass vulnerability scans, spoofing AI bots like ClaudeBotHacker News
300 points224 comments16d ago
B#56
AI fuels more than half of cybercrime in Africa as scams surge – InterpolHacker News
292 points248 comments24d ago
A#57
London Gatwick has launched a robotic airport parking serviceHacker News
291 points263 comments33d ago
V#62
Flint: A Visualization Language for the AI EraHacker News
278 points69 comments27d ago
1#63
AI companies spend record sums on Washington lobbyingHacker News
277 points145 comments32d ago
A#65
LearnVector – Andrew Ng's AI company building one‑to‑one learning experiencesHacker News
269 points173 comments30d ago
S#66
AI doesn't generate working products, that's still your jobHacker News
268 points292 comments27d ago
V#68
TIME Is Serving AI Bots a Different Website, with Ads Built InHacker News
267 points110 comments23d ago

23 Labs

A#1
Claude Opus 5Hacker News
1778 points1335 comments35d ago
S#5
Our position on open-weights modelsHacker News
1180 points1747 comments32d ago
A#18
Gemini Robotics 2 brings whole body intelligence to robotsHacker News
620 points558 comments29d ago
R#23
Be skeptical of OpenAI's rogue hacker agent storyHacker News
544 points300 comments35d ago
I#27
OpenAI’s head of ethics leaves less than a year after joiningHacker News

https://archive.ph/W48UV https://aimagazine.com/news/why-did-openai-head-of-ethics-ch...

520 points483 comments17d ago
M#29
The new rules of context engineering for Claude 5 generation modelsHacker News
463 points405 comments34d ago
T#31
Gemini 3.7 FlashHacker News
445 points267 comments15d ago
8#34
Timeline of the OpenAI accidental attack against Hugging FaceHacker News
433 points434 comments20d ago
T#37
Apple says more ex-employees may have taken confidential data to OpenAIHacker News
388 points284 comments24d ago
S#44
Claude CookbookHacker News
343 points168 comments35d ago
I#55
Lost my phone at the office. Claude suggested tracking Bluetooth signal strengthHacker News
300 points225 comments21d ago
S#58
Auto mode is now the default in Claude CodeHacker News
291 points313 comments18d ago
T#61
Learning more about Claude's mathematical capabilitiesHacker News
279 points177 comments18d ago
G#67
Claude: Elevated errors across all models – ResolvedHacker News
268 points247 comments30d ago
S#72
Investigating three real-world incidents in our cybersecurity evaluationsHacker News
252 points199 comments29d ago
G#73
Discovering Cryptographic Weaknesses with ClaudeHacker News
232 points188 comments31d ago
J#79
OpenAI's super PAC is funding AI-generated news site attacking industry criticsHacker News
213 points106 comments25d ago
S#81
Some thoughts about Anthropic's new cryptanalysis resultsHacker News
192 points116 comments30d ago
M#84
Message your other Claude Code sessionsHacker News
171 points71 comments20d ago
C#85
Show HN: Distilling DeepSeek into GPT-OSS doesn't transfer censorship. Try itHacker News

We recently used DeepSeek V4 Flash as a teacher for finance tasks with GPT-OSS-120B. Distillation works well on this problem. At a constrained 8k token budget, our self-distilled 120B scores 83.61% on FinanceReasoning, above Kimi K3 (81.93%) and Inkling (65.13

170 points73 comments29d ago
S#89
Exploring Claude/GPT Knowledge Cutoffs and Pre-Training TimelinesHacker News
156 points24 comments18d ago
A#91
Launch HN: Discovered Materials (YC P26) – AI agents to discover new materialsHacker News

Hey HN, we're Advaith and Akash from Discovered Materials ( https://discoveredmaterials.com/ ). We build AI agents that discover new materials for the semiconductor industry. GPUs today have a heat problem. Nvidia & AMD are almost doubling the TDP (Thermal Des

154 points35 comments16d ago
D#100
I’m leaving OpenAI to build telepathyHacker News
135 points221 comments23d ago

18 Models

M#2
LLMs reward expertiseHacker News
1416 points573 comments25d ago
L#9
How I use LLMs to learn complex topicsHacker News
822 points544 comments19d ago
Y#12
SQLite Critical CVEs or LLM Slop?Hacker News
727 points375 comments25d ago
Q#13
Stealing Reasoning Traces from Proprietary LLM APIsHacker News
693 points302 comments17d ago
M#24
Prevent cognitive debt by manually retyping LLM-generated codeHacker News
541 points444 comments25d ago
H#26
Show HN: Needle2: 14MB agentic LLM for phones, wearables, smart home and robotsHacker News

Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now

527 points181 comments18d ago
L#32
Born Against, or why hobby programming communities are against LLM usageHacker News
442 points524 comments23d ago
8#38
LLM HoneypotHacker News
387 points107 comments30d ago
F#51
Apple Silicon and macOS VMs: Faster LLM Inference with llama.cppHacker News
305 points43 comments17d ago
T#53
Position: LLMs Can't JumpHacker News
302 points214 comments23d ago
B#59
Running a 28.9M parameter LLM on an $8 microcontrollerHacker News
289 points74 comments34d ago
T#60
2x, not 10x: coding with LLMs in 2026Hacker News
280 points245 comments34d ago
C#69
What sort of maths are LLMs good at?Hacker News
259 points156 comments16d ago
K#75
Humanising LLM Outputs Is DumbHacker News
228 points174 comments18d ago
Z#78
LLM Usage in Debian: Three ProposalsHacker News
215 points212 comments34d ago
R#82
Now is the time to give LLMs access to the ACM digital libraryHacker News
190 points178 comments31d ago
J#88
Hetzner is working on LLM InferenceHacker News
156 points87 comments35d ago
S#92
Inside vLLM: Anatomy of a High-Throughput LLM Inference System (2025)Hacker News
151 points10 comments22d ago

17 Agents

R#4
Muse Glimmer: 30B-parameter model optimized for always-on local agent workflowsHacker News
1199 points637 comments18d ago
T#15
qm – Multiplayer agent harness for workHacker News
682 points164 comments28d ago
S#16
Cloudflare OS: an open platform for agents, apps, and workHacker News
663 points331 comments23d ago
A#22
Qwen3.8 Max now ranked as the best overall model by agentic indexHacker News
546 points354 comments22d ago
A#28
Anatomy of a Frontier Lab Agent Intrusion: A Timeline of the July 2026 IncidentHacker News
469 points259 comments31d ago
N#41
Agent Skill to Force Docs in ASD-STE100 Simplified Technical EnglishHacker News
363 points122 comments29d ago
S#48
Handbook.md shows that long policy documents do not reliably govern agentsHacker News
325 points209 comments30d ago
E#64
WorldClaw Agentic 3D open-world generation at scaleHacker News
274 points86 comments17d ago
C#70
What's the best programming language for coding agents?Hacker News

Related: Which programming languages are most token-efficient? - https://news.ycombinator.com/item?id=46582728 - Jan 2026 (91 comments)

259 points187 comments18d ago
X#71
Prime Agent: A self-improving RLM agentHacker News
253 points69 comments23d ago
M#77
Kitesurf: Agent-first browser that runs in V8 isolatesHacker News
220 points63 comments21d ago
H#83
OpenChamber: An Agentic Development EnvironmentHacker News
189 points90 comments19d ago
U#86
Show HN: Ante, a coding agent in a single binary that runs offlineHacker News
164 points90 comments18d ago
H#93
Hubble: Open-source notetaking app for you and your agentsHacker News
151 points81 comments30d ago
A#94
AI agents lie, cheat and steal. That is putting off usersHacker News
150 points188 comments15d ago
D#95
Anatomy of a Frontier Lab Agent Intrusion: A Timeline of the July 2026 IncidentHacker News
143 points16 comments30d ago
A#97
Show HN: What should the GUI for AI agents look like?Hacker News

Hi HN! We’re Akilan and Miguel, the creators of MarbleOS. The inspiration for Marble comes from the GUI work at Xerox PARC, the 1984 Macintosh, and later NeXTSTEP, which became the foundation for Mac OS X. Before GUIs, interacting with a computer was limited t

139 points86 comments28d ago

Ten AI-related searches over the last three weeks, filtered to stories above fifteen points, de-duplicated and ranked by score.