102 stories tagged llms, newest first.
169 points by drewsberry | 203 comments
Blog post surveying a week of critical software vulnerabilities, including PHP2Shell RCEs found with AI, and discussing whether LLM-driven exploitation is the new normal.
23 points by paulorlando | 9 comments
Blog essay on mathematicians' emotional reactions to AI-generated proofs, framing it as the end of math's romantic era and comparing it to past industrial disruptions.
18 points by gopiandcode | 7 comments
Basis blog post describing using LLMs to write a formal semantics for a small language and Lean to prove its interpreter correct down to RISC-V compiled code.
63 points by dajonker | 61 comments
Blog post arguing that Ruby on Rails and frameworks generally still make sense even as LLMs make writing code cheap.
44 points by sanketsaurav | 39 comments
Blog essay arguing that AI coding tools like Claude Code have ended the rationing of engineering effort, and software makers haven't yet adapted.
25 points by guitarlimeo | 32 comments
Ask HN discussion asking whether senior engineers' architecture and design roles will survive as LLM coding agents improve through reinforcement learning.
42 points by mfrw | 35 comments
A post on Niko Matsakis's baby steps blog reflecting on what LLM coding agents mean for the Rust language and how Rust can benefit agentic development.
16 points by speckx | 14 comments
Blog post on Nisheeth Vishnoi's site 'Algorithms, Nature, and Society' asking who taught AI systems mathematics, illustrated with handwritten graph theory notes.
128 points by DeepLogin | 80 comments
A blog post by Micah Lee arguing that people should refuse to use or support OpenAI and Anthropic.
662 points by albertzeyer | 963 comments
A post by Andrej Karpathy on X arguing that the gap in understanding of LLM capability is widening into a funnel, from non-users to professional users of coding agents.
27 points by rizsyed1 | 32 comments
Blog post by Iris Meredith on deadSimpleTech arguing that AI and LLM adoption is not inevitable, calling the claim a thought-terminating cliché.
68 points by mtlynch | 47 comments
Michael Lynch blog post arguing that AI coding agents remain buggy and primitive despite capable underlying LLMs, reflecting on his experience since February 2025.
86 points by vaughands | 44 comments
Personal blog post by Vaughan Hilts describing how using LLMs reduced his typing and appears to have eased his long-standing repetitive strain injury.
52 points by colinarms | 38 comments
A personal blog essay on how LLMs make it worthwhile to build small personal tools that previously wouldn't justify the effort.
24 points by airhangerf15 | 34 comments
A 404 Media article about a project that 'tortures' LLMs in a simulated robot prison and the online debate it sparked over AI sentience and welfare.
40 points by dblack12705 | 14 comments
Substack post by Dylan Black examining the calibration of Jev, TypeSafe's "System One" classifier model, contrasted with LLMs as System Two models.
41 points by leopoldj | 42 comments
MIT Technology Review opinion piece by Thore Graepel arguing that LLMs don't truly reason, ten years after AlphaGo's match against Lee Sedol.
206 points by benbreen | 64 comments
Res Obscura Substack post by Benjamin Breen on using the Opus 5.5 AI model to find a previously unknown eyewitness account of the dodo.
92 points by ArchAndStarch | 68 comments
Instructor Jeff Turkstra's retrospective on an AI/LLM cheating incident in his Spring 2026 CS 240 Programming in C course and how it was handled.
300 points by allisdust | 269 comments
A blog post arguing about the Western vs. Chinese AI lab race, discussing Anthropic's claims that Chinese labs are distilling its models.
291 points by signa11 | 307 comments
A Haskell Discourse forum post discussing how programmers can keep enjoying programming amid the rise of LLM coding tools and AI burnout.
268 points by azhenley | 257 comments
Christian Kästner's Substack post describing how he redesigned course assessments after AI agents could complete all his homework assignments.
160 points by benbreen | 38 comments
A Res Obscura Substack post by historian Benjamin Breen arguing AI labs should fund historical research, with examples of using frontier LLMs to trace alchemy and decode 17th-century letters.
37 points by aquastorm | 39 comments
An I Programmer opinion piece by Mike James arguing that AI has no intent or motivation, so fears of AI going rogue are misrepresentations.
326 points by teoruiz | 214 comments
Blog post on jyn.dev arguing LLM token prices are falling by orders of magnitude, so quality and access, not token cost, will limit AI use.
105 points by gmays | 42 comments
Nathan Lambert's Interconnects post expanding his Congressional testimony on the state of open-weight AI models amid U.S.-China competition.
311 points by dimonomid | 438 comments
Blog post by Alexandru Nedelcu arguing that AI lacks wisdom and that relying on it for coding, skipping code reviews and reading code, erodes developers' own judgment.
426 points by jaykru | 552 comments
Personal blog essay by Jay Kruer arguing LLMs remain overhyped despite headline results like Navier-Stokes, citing poor generalization, reward hacking, and specification costs.
1171 points by meredydd | 1120 comments
A declaration by mathematicians arguing that AI companies' use of math problems as benchmarks harms mathematics and the mathematical community, inviting endorsements.
490 points by larsfaye | 479 comments
Blog essay by Lars Faye arguing that reliance on AI coding tools removes the friction needed to develop long-term programming expertise.
566 points by bilsbie | 645 comments
A Paul Graham post on X saying that at 17 he would learn to build LLMs from scratch rather than start a startup, with replies including Yann LeCun.
592 points by Jach | 442 comments
Dan Luu essay arguing that LLM coding agents make specialized performance optimization cheap, illustrated with optimizing a regex engine for specific workloads.
763 points by laurentiurad | 489 comments
Personal blog post by Laurentiu Raducu describing how he uses LLMs to learn complex topics, including building an interactive chip-making game called ChipTycoon.
761 points by senko | 446 comments
Blog post by Senko Rašić arguing that the claim "code was never the hard part" amid LLM coding hype is an insult to programmers.
1278 points by MaxMussio | 527 comments
Blog post by Sean Goedecke arguing that the most important skill in prompting LLMs is expertise in the domain being prompted for.
702 points by pchm | 526 comments
Blog essay arguing that despite hype about AI solving coding and an 'Agentic Era', software quality keeps getting worse.
387 points by droidjj | 209 comments
Simon Willison blog post on Moonshot AI's 2.8-trillion-parameter Kimi K3 model, its benchmarks and pricing, and his pelican-on-a-bicycle test results.
409 points by therepanic | 253 comments
A blog post by George Hotz (geohot) saying he is enthusiastic about LLMs and AI progress but dislikes fear-driven and exaggerated AI hype.
676 points by lumpa | 556 comments
Simon Willison blog post describing Claude Fable 5 as relentlessly proactive, illustrated with an example of fixing a scrollbar glitch in Datasette Agent.
1613 points by jjfoooo4 | 477 comments
Blog post arguing that people forwarding AI-generated text to colleagues should first digest it and demonstrate their own effort before asking for others' attention.
1029 points by poisonfountain | 967 comments
A personal blog post by a software engineer with 10 years of experience reflecting on how LLMs are eroding their career and their uncertainty about what to do.
646 points by andrehacker | 1029 comments
Ask HN discussion thread asking users to describe the moment they realized generative AI models were more capable than they had assumed.
457 points by kostaj | 314 comments
Lenz research report finding that five frontier LLMs disagree on the verdict for 67% of 1,000 real-world fact-check claims.
347 points by razin | 280 comments
A blog post by George Hotz arguing that adopting AI coding agents in software development is a costly mistake because their output is subtly broken.
1366 points by dmarcos | 572 comments
Tweet by Andrej Karpathy announcing he has joined Anthropic to return to LLM R&D, while planning to resume his education work later.
717 points by yakkomajuri | 542 comments
Simon Willison's annotated slides from a five-minute PyCon US 2026 lightning talk summarizing the last six months of LLM developments.
380 points by movis | 624 comments
Sean Goedecke's blog post arguing that because AI use reduces learning on the job, software engineering may no longer be a lifelong career.
728 points by e12e | 823 comments
Simon Willison blog post recapping a Heavybit podcast appearance, reflecting that vibe coding and responsible agentic engineering are converging in his own work.
575 points by indigodaddy | 348 comments
Simon Willison's blog post reviewing DeepSeek's V4 preview models, Pro and Flash, 1M-context MoE open-weights LLMs under MIT license, and their pricing.
618 points by lumpa | 401 comments
Simon Willison note on Zig's strict anti-LLM contribution policy and its rationale, noting Bun's fork of Zig and its AI-assisted compiler improvements.
337 points by pretext | 186 comments
Simon Willison's blog post comparing Anthropic's published Claude system prompts for Opus 4.6 and Opus 4.7, using a git diff to highlight the changes.
372 points by kiyanwang | 238 comments
Blog post by Viktor Cessan analyzing the financial costs of software teams, why organizations lack visibility into them, and how LLMs change the picture.
566 points by stevekrouse | 419 comments
Steve Krouse blog post arguing that code will not be replaced by AI-written specs, since a sufficiently detailed spec is itself code.
440 points by indigodaddy | 423 comments
A personal blog post by Stavros describing how he uses LLMs to write software and why he values making things over programming itself.
760 points by simonw | 333 comments
Simon Willison blog post on Qwen 3.5 open weight models and the departure of lead researcher Junyang Lin from Alibaba's Qwen team after a reorg.
634 points by ColinWright | 737 comments
Nolan Lawson's blog post lamenting how AI coding tools are reducing programmers to code reviewers and mourning the loss of the programming craft.
432 points by simonw | 305 comments
Simon Willison's blog post describing undocumented upgrades to ChatGPT's code container: running Bash, Node.js and other languages, pip/npm installs via proxy, and file downloads.
823 points by simonw | 451 comments
Simon Willison's annual blog post reviewing major LLM trends of 2025, including reasoning models, coding agents, Chinese open weight models, and Gemini.
777 points by simonw | 599 comments
Simon Willison blog post arguing developers must prove their code works through manual and automated testing before submitting it, especially when using LLM tools.
376 points by martinald | 584 comments
Blog post by Martin Alderson arguing that agentic AI coding has dramatically cut the cost of building software and will transform the industry by 2026.
636 points by steveklabnik | 249 comments
Oxide Computer's public RFD 576 by Bryan Cantrill describing how the company approaches using large language models in its work.
1064 points by ctoth | 953 comments
Dwarkesh Podcast episode interviewing Andrej Karpathy, who argues it will take about a decade to work through the problems with AI agents.
576 points by janpio | 612 comments
Simon Willison blog post proposing the term "vibe engineering" for experienced engineers using LLMs and coding agents while staying accountable for their code.
618 points by chrisloy | 377 comments
Blog post by Chris Loy arguing that AI coding tools speed up typing code but risk neglecting the problem-solving that is the real work of software development.
526 points by dbalatero | 319 comments
A Substack essay by Mike Judge arguing that claims of massive AI coding productivity gains aren't supported by evidence such as a surge in new software releases.
450 points by tosh | 304 comments
Short blog post by Simon Willison proposing the analogy that an LLM is a lossy encyclopedia, and advising users to supply correct examples rather than rely on recall.
378 points by floverfelt | 347 comments
Martin Fowler's blog post sharing scattered thoughts on the state of LLMs and AI and their effect on software development.
324 points by maloga | 211 comments
Blog post by Mariano Gappa comparing building a card game in Go over three months without LLMs versus another in three days with LLM assistance.
321 points by srid | 196 comments
A Zed blog post by Conrad Irwin arguing that LLMs cannot really build software, drawing on his experience of what effective software engineers do.
330 points by ingve | 189 comments
Blog post by Yossi Kreinin arguing that large language models are not world models, illustrated with examples like LLMs losing track of chess board state.
377 points by tosh | 365 comments
Simon Willison blog post on OpenAI retiring GPT-4o and other older models in ChatGPT consumer apps without notice when GPT-5 launched.
325 points by jjfoooo4 | 183 comments
Blog post on the Elroy site arguing AI reshapes learning curves by helping beginners reach competence faster, while mastery and large-codebase work remain hard.
320 points by eternalreturn | 165 comments
Blog post by Hazel Weakly arguing the industry builds AI tools backwards, drawing on research about how humans learn and transfer knowledge.
356 points by Dachande663 | 204 comments
Blog post by Utkarsh Kanwat arguing that the hype around fully autonomous AI agents is unrealistic and describing what actually works in production agent systems.
1607 points by SwoopsFromAbove | 1502 comments
Blog post by Tom Renner titled "The sound of inevitability" arguing against framing widespread LLM adoption as inevitable.
453 points by zdw | 377 comments
Blog post on ADD / XOR / ROL arguing LLMs should be viewed as mathematical functions mapping paths through embedding space rather than human-like agents.
448 points by bilsbie | 240 comments
Jack Morris essay arguing that major AI breakthroughs, including LLMs, came from new datasets rather than new ideas or methods.
![Screenshot of youtube.com: Andrej Karpathy: Software in the era of AI [video]](https://hn.urlbox.com/shots/renders/2025/6/19/02a9b2f5-6352-4dd8-82bd-7e40d4498927.jpg)
1382 points by sandslash | 757 comments
YouTube video from Y Combinator of Andrej Karpathy's talk "Software Is Changing (Again)" about how AI and LLMs are reshaping software.
466 points by bumbledraven | 261 comments
Blog post by David Crawshaw on using LLM-based agents for programming, defining an agent as a loop around an LLM that executes commands.
866 points by swyx | 216 comments
Simon Willison's writeup of his AI Engineer World's Fair keynote reviewing six months of LLM releases, compared by having each model draw a pelican on a bicycle.
526 points by longwave | 604 comments
Blog post by antirez arguing human programmers still outperform LLMs, illustrated by debugging a Redis Vector Sets corruption issue.
588 points by bradgessler | 367 comments
Dustin Curtis's blog essay reflecting on how AI and LLMs have made his own writing, coding, and creative thinking feel pointless.
335 points by minimaxir | 183 comments
Max Woolf's blog post describing how he, as an experienced data scientist and LLM tool developer, actually uses generative LLMs only sparingly in his work.
336 points by NotInOurNames | 293 comments
A Substack post by Addy Osmani on avoiding skill atrophy when relying on AI tools; the captured page shows only an error message.
388 points by dhorthy | 57 comments
GitHub repository by HumanLayer documenting twelve principles for building reliable, production-grade LLM-powered agent applications, modeled on 12 Factor Apps.
611 points by ksec | 1030 comments
A post on X (Twitter) by Sabine Hossenfelder questioning why people remain bullish about large language models; the page did not load its content.
355 points by finding_theta | 158 comments
Blog post by Chiraag Gohel comparing how well LLMs perform exploratory data analysis and whether they notice a hidden 'gorilla' pattern in a dataset.
773 points by cryptophreak | 395 comments
Blog post by Daniel Delaney arguing that chat is a poor user interface pattern for AI-powered software development tools.
392 points by twapi | 307 comments
Simon Willison blog post quoting Anthropic's job application form asking applicants not to use AI assistants when applying.
400 points by ingve | 56 comments
Latent Space post listing 50 papers, models, and blogs across 10 AI engineering fields such as LLMs, benchmarks, prompting, RAG, agents, and finetuning.