Top all-time · #1581
Cognition
Company blog · cognition.ai/blogaiprogramming
Cognition is the company blog at cognition.ai, writing about ai and programming. Since 2024, 45 different Hacker News readers have linked it 59 times, most often “FrontierCode”. No feed was found, so read it on the site itself.
Cognition is the applied AI lab behind Devin, the first autonomous software engineer. We build agents and tools that expand human capacity.
Most-linked posts?
- FrontierCode5 people · 5 mentions · first linked 2026-06-08
- Don't Build Multi-Agents4 people · 5 mentions · first linked 2025-06-12
- Rebuilding Devin for Claude Sonnet 4.5: Lessons and Challenges3 people · 3 mentions · first linked 2025-09-29
- Blockdiff: We built our own file format for VM disk snapshots3 people · 3 mentions · first linked 2025-06-23
- SWE-Grep and SWE-Grep-Mini: RL for Fast Multi-Turn Context Retrieval3 people · 3 mentions · first linked 2025-10-16
- A review of OpenAI o1 and how we evaluate coding agents2 people · 3 mentions · first linked 2024-09-12
- Agent Trace: Capturing the Context Graph of Code2 people · 2 mentions · first linked 2026-01-30
- Estimating the Productivity of an Autonomous AI Software Engineer2 people · 2 mentions · first linked 2026-06-10
- Using Devin to Recover from the CrowdStrike Outage2 people · 2 mentions · first linked 2024-07-20
- Cognition Releases SWE-1.5: Near-SOTA Coding Performance at 950 tok/s2 people · 2 mentions · first linked 2025-10-29
What Hacker News says?
> Also... telling that they choose to benchmark against CodeForces rather than SWE-bench. They also worked with Devin to benchmark it on Devin's internal benchmarks, where it's twice as good as GPT-4o: https://x.com/cognition_labs/status/1834292718174077014 https://www.cognition.ai/blog/evaluating-coding-agents
Comment linking /blog/evaluating-coding-agents
on “Learning to Reason with LLMs”Submitted Cognition (Devin AI) to Acquire Windsurf as a story · 502 points
HN discussionCognition (Devin) got early access. Interesting write-up: https://www.cognition.ai/blog/evaluating-coding-agents
Comment linking /blog/evaluating-coding-agents
on “Learning to Reason with LLMs”Submitted Codemaps: Understand Code, Before You Vibe It as a story · 315 points
HN discussionYes, but at the same time having the 1 million context enabled is nice because the model is aware that they have more context left and actually perform better. [0] [0] https://cognition.ai/blog/devin-sonnet-4-5-lessons-and-chall...
Comment linking /blog/devin-sonnet-4-5-lessons-and-challenges
on “Claude Code 2.0”Submitted Devin is now generally available as a story · 155 points
HN discussionhey we been shipping!! https://cognition.ai/blog/swe-1-5 https://cognition.ai/blog/swe-grep https://cognition.ai/blog/devin-agent-preview-sonnet-4-5
Comment linking /blog/devin-agent-preview-sonnet-4-5
on “Codemaps: Understand Code, Before You Vibe It”Submitted Don't Build Multi-Agents as a story · 123 points
HN discussionI think both Cursor and Cognition and going in the same direction of SWE-grep[0]. SWE-grep was able to hit ~700tokens/s and Cursor ~300token/s, hard to compare the precision/recall and cost effectiveness though, considering SWE-grep also adopted a "hack" of running it on Cerebras. I'm trying to kickstart a RL-based code search project called "op-grep" here[1], still pretty early, but looking for collaborators! [0]:…
Comment linking /blog/swe-grep
on “Composer: Building a fast frontier model with RL”Submitted SWE-Grep and SWE-Grep-Mini: RL for Fast Multi-Turn Context Retrieval as a story · 97 points
HN discussionIt's a reference to Devin, one of the earlier (and most hyped) "autonomous" ai-agent-based software devs that it attempts to replicate/match in the open. Your interestingly different ire would be better-directed at the original project. https://www.cognition.ai/blog/introducing-devin Previous discussions on that fwiw include: https://news.ycombinator.com/item?id=39679787
Comment linking /blog/introducing-devin
on “OpenDevin: An Open Platform for AI Software Developers as Generalist Agents”Submitted Blockdiff: We built our own file format for VM disk snapshots as a story · 91 points
HN discussionAh. Thx. Blogpost for others: https://cognition.ai/blog/swe-1-5 Takeaway is that this is sonnet-ish model at 10x the speed.
Comment linking /blog/swe-1-5
on “Cerebras Code now supports GLM 4.6 at 1000 tokens/sec”Submitted A review of OpenAI o1 and how we evaluate coding agents as a story · 34 points
HN discussionIt is, but there's a direct tension here between security and capabilities. It's hard to do useful things with private data without opening up prompt injection holes. And there's a huge demand for this kind of product. Agents also typically work better when you combine all the relevant context as much as possible rather than splitting out and isolating context. See: https://cognition.ai/blog/dont-build-multi-agents…
Comment linking /blog/dont-build-multi-agents
on “How to stop AI's "lethal trifecta"”Submitted Devin Review: AI to Stop Slop as a story · 26 points
HN discussionThis link redirects to https://cognition.ai/blog/swe-grep now?
Comment linking /blog/swe-grep
on “SWE-Grep and SWE-Grep-Mini: RL for Fast Multi-Turn Context Retrieval”Submitted Cognition Releases SWE-1.5: Near-SOTA Coding Performance at 950 tok/s as a story · 11 points
HN discussionhttps://cognition.ai/blog/frontier-code (disclaimer - was on the team - but also we covered swebench pro/deepswe issues in here as well.)
Comment linking /blog/frontier-code
on “Separating signal from noise in coding evaluations”Submitted Show HN: Blockdiff – How we built our own file format for VM disk snapshots as a story · 11 points
HN discussionAgreed. In fact, I thought this was an iteration on the previously announced "Devin" project until I realised that they have different spelling. https://www.cognition.ai/introducing-devin
Comment linking /introducing-devin
on “Devon: An open-source pair programmer”Submitted Show HN: Kevin-32B – how to do multi-turn RL on writing CUDA kernels as a story · 7 points
HN discussionAlso on the HN homepage: https://cognition.ai/blog/frontier-code Not quite ASCII but pretty close!
Comment linking /blog/frontier-code
on “Show HN: Performative-UI – a react component library of design tropes”Submitted Windsurf Codemaps: Understand Code, Before You Vibe It as a story · 6 points
HN discussionCOBOL migration is one of Devin's advertised capabilities: https://docs.devin.ai/use-cases/examples/cobol-modernization https://cognition.ai/blog/infosys-cognition
Comment linking /blog/infosys-cognition
on “Ask HN: COBOL devs, how are AI coding affecting your work?”Submitted Devin 2.0 as a story · 6 points
HN discussionOn the new FrontierCode [1] benchmark (ie graded from an OSS maintainer's perspective of "would I merge this code?") - Opus 4.7 xhigh: 5.2% - Opus 4.8 xhigh: 13.4% - Fable 5 xhigh: 29.3% Seems like a huge jump. [1] https://cognition.ai/blog/frontier-code
Comment linking /blog/frontier-code
on “Claude Fable 5”Submitted Don't Build Multi-Agents as a story · 5 points
HN discussionIt might be clearer to link the original blog post ( https://cognition.ai/blog/kevin-32b ) rather than the Hugging Face model card. The charts and formulas are attractive, but the CUDA snippet feels quite elementary.
Comment linking /blog/kevin-32b
on “Devin's First Open Source Model Beats O3”Submitted Devin makers: you shouldn't build multi-agent systems as a story · 5 points
HN discussion
