{"slug": "devin", "name": "Devin", "category": "ai-coding", "category_name": "AI coding tools", "description": "Autonomous AI software engineering agent from Cognition that writes, tests, and submits code changes.", "recommenders": 5, "commenters": 14, "warners": 4, "score": 2.384, "rising": 3.67, "first": "2025-12-03", "latest": "2026-10-16", "why": "Liked as a cloud-hosted, team-managed coding agent that works in its own sandboxed environment and opens PRs, and several commenters say it has improved a lot.", "why_is_quote": false, "has_verdict": true, "sentiment": {"recommend": 5, "neutral": 3, "compare": 2, "warn": 4}, "spark": [0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 4, 0], "listed": true, "loved": ["Team-level agent", "Sandboxed environment"], "caveat": "Weak results", "short_quote": "We use https://devin.ai for this and it works very well.", "icon": true, "domain": "devin.ai", "github_repo": null, "official_url": "https://devin.ai", "pricing": "paid", "aliases": ["Devin"], "verdict": [{"text": "Liked as a cloud-hosted, team-managed coding agent that works in its own sandboxed environment and opens PRs,", "cites": [{"id": 794424784, "item_id": 49651549, "url": "https://news.ycombinator.com/item?id=49651549", "by": "deet", "date": "2026-09-10", "stance": "recommend", "quote": "at this point is one of the best cloud-hosted, team-managed coding agents", "excerpt": "As others have mentioned, it's really matured a lot and at this point is one of the best cloud-hosted, team-managed coding agents, when factoring overall UX, testing and QA lifecycle via its sandboxes, and its ability to be controlled with an API. We use it quite heavily. It's coming from a different starting place than Claude Code or Codex are as individually controlled single-developer tools. Devin has been more pe", "context": "team using it heavily", "use_case": "team-managed autonomous coding agent", "switched_from": "", "story": "Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra"}, {"id": 745803440, "item_id": 46612715, "url": "https://news.ycombinator.com/item?id=46612715", "by": "cloudking", "date": "2026-01-13", "stance": "recommend", "quote": "We use https://devin.ai for this and it works very well.", "excerpt": "We use https://devin.ai for this and it works very well. Devin has it's own virtual environment, IDE, terminal and browser. You can configure it to run your application and connect to whatever it needs. Devin can modify the app, test changes in the browser and send you a screen recording of the working feature with a PR.", "context": "", "use_case": "background coding agent", "switched_from": "", "story": "Why We Built Our Own Background Agent"}]}, {"text": " and several commenters say it has improved a lot.", "cites": [{"id": 794424784, "item_id": 49651549, "url": "https://news.ycombinator.com/item?id=49651549", "by": "deet", "date": "2026-09-10", "stance": "recommend", "quote": "at this point is one of the best cloud-hosted, team-managed coding agents", "excerpt": "As others have mentioned, it's really matured a lot and at this point is one of the best cloud-hosted, team-managed coding agents, when factoring overall UX, testing and QA lifecycle via its sandboxes, and its ability to be controlled with an API. We use it quite heavily. It's coming from a different starting place than Claude Code or Codex are as individually controlled single-developer tools. Devin has been more pe", "context": "team using it heavily", "use_case": "team-managed autonomous coding agent", "switched_from": "", "story": "Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra"}, {"id": 793934832, "item_id": 49620927, "url": "https://news.ycombinator.com/item?id=49620927", "by": "vvern", "date": "2026-09-09", "stance": "recommend", "quote": "I used to be a hater, but it\u2019s kinda good now", "excerpt": "I used to be a hater, but it\u2019s kinda good now", "context": "", "use_case": "", "switched_from": "", "story": "Cognition (Devin) raises $2B at $48B valuation"}]}, {"text": " The most common complaint is that it struggles with anything beyond trivial work,", "cites": [{"id": 794413664, "item_id": 49650854, "url": "https://news.ycombinator.com/item?id=49650854", "by": "SaltyBackendGuy", "date": "2026-09-10", "stance": "warn", "quote": "It sucked the second time also...", "excerpt": "> And yes, I tried again This made me laugh a bit. I was forced to do an evaluation of their shit product twice due to being backed by the same PE firm; \"take a look at it again, it's much better now\". It sucked the second time also...", "context": "forced to evaluate it twice due to PE firm backing", "use_case": "AI coding agent evaluation", "switched_from": "", "story": "Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra"}, {"id": 738255361, "item_id": 46140960, "url": "https://news.ycombinator.com/item?id=46140960", "by": "rr808", "date": "2025-12-03", "stance": "warn", "quote": "most people can't even get the tools to do anything useful more than trivial things", "excerpt": "We have these weekly rah rah AI meetings where we swap tips on what we've achieved with copilot and devin. Mostly crickets but everyone is talking with lots of enthusiasm. Its starting to get silly though now, most people can't even get the tools to do anything useful more than trivial things we used to see on stack overflow.", "context": "employee at a company with weekly AI meetings", "use_case": "AI coding at work", "switched_from": "", "story": "Everyone in Seattle hates AI"}]}, {"text": " and some find Claude Code or Codex better.", "cites": [{"id": 794391312, "item_id": 49649457, "url": "https://news.ycombinator.com/item?id=49649457", "by": "hightrix", "date": "2026-09-10", "stance": "recommend", "quote": "It's a great product compared to Copilot.", "excerpt": "My company uses it. We have a bunch of seats in an enterprise plan and have been using it for 9 months or so. It's a great product compared to Copilot. It is also the first AI tool I used heavily outside of creating random images or one off questions. I'm now using all three, Devin, Claude, Codex. I'm finding Claude and Codex to be much better. One of my biggest gripes is that the web client and desktop client for De", "context": "company with enterprise plan seats", "use_case": "AI coding at company", "switched_from": "", "story": "Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra"}, {"id": 776941568, "item_id": 48558848, "url": "https://news.ycombinator.com/item?id=48558848", "by": "jatora", "date": "2026-06-16", "stance": "warn", "quote": "I cant believe people still use Devin.", "excerpt": "I cant believe people still use Devin. This is just filling the niche of vibe PR'ing? Not even vibe coding? Another level removed from the codebase than even claude code/codex are?", "context": "", "use_case": "", "switched_from": "", "story": "SpaceX to buy Cursor AI coding agent operator Anysphere for $60B"}]}], "praise": [{"label": "Team-level agent", "summary": "Seen as a strong cloud-hosted, team-managed agent that works autonomously and can be controlled via API.", "quotes": [{"id": 794424784, "item_id": 49651549, "url": "https://news.ycombinator.com/item?id=49651549", "by": "deet", "date": "2026-09-10", "stance": "recommend", "quote": "at this point is one of the best cloud-hosted, team-managed coding agents", "excerpt": "As others have mentioned, it's really matured a lot and at this point is one of the best cloud-hosted, team-managed coding agents, when factoring overall UX, testing and QA lifecycle via its sandboxes, and its ability to be controlled with an API. We use it quite heavily. It's coming from a different starting place than Claude Code or Codex are as individually controlled single-developer tools. Devin has been more pe", "context": "team using it heavily", "use_case": "team-managed autonomous coding agent", "switched_from": "", "story": "Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra"}, {"id": 794391312, "item_id": 49649457, "url": "https://news.ycombinator.com/item?id=49649457", "by": "hightrix", "date": "2026-09-10", "stance": "recommend", "quote": "It's a great product compared to Copilot.", "excerpt": "My company uses it. We have a bunch of seats in an enterprise plan and have been using it for 9 months or so. It's a great product compared to Copilot. It is also the first AI tool I used heavily outside of creating random images or one off questions. I'm now using all three, Devin, Claude, Codex. I'm finding Claude and Codex to be much better. One of my biggest gripes is that the web client and desktop client for De", "context": "company with enterprise plan seats", "use_case": "AI coding at company", "switched_from": "", "story": "Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra"}]}, {"label": "Sandboxed environment", "summary": "Has its own VM, IDE, terminal and browser to run, test and record changes before sending a PR.", "quotes": [{"id": 745803440, "item_id": 46612715, "url": "https://news.ycombinator.com/item?id=46612715", "by": "cloudking", "date": "2026-01-13", "stance": "recommend", "quote": "We use https://devin.ai for this and it works very well.", "excerpt": "We use https://devin.ai for this and it works very well. Devin has it's own virtual environment, IDE, terminal and browser. You can configure it to run your application and connect to whatever it needs. Devin can modify the app, test changes in the browser and send you a screen recording of the working feature with a PR.", "context": "", "use_case": "background coding agent", "switched_from": "", "story": "Why We Built Our Own Background Agent"}, {"id": 794424784, "item_id": 49651549, "url": "https://news.ycombinator.com/item?id=49651549", "by": "deet", "date": "2026-09-10", "stance": "recommend", "quote": "at this point is one of the best cloud-hosted, team-managed coding agents", "excerpt": "As others have mentioned, it's really matured a lot and at this point is one of the best cloud-hosted, team-managed coding agents, when factoring overall UX, testing and QA lifecycle via its sandboxes, and its ability to be controlled with an API. We use it quite heavily. It's coming from a different starting place than Claude Code or Codex are as individually controlled single-developer tools. Devin has been more pe", "context": "team using it heavily", "use_case": "team-managed autonomous coding agent", "switched_from": "", "story": "Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra"}]}, {"label": "Much improved", "summary": "Former skeptics and long-time users say it has matured and is now fairly good.", "quotes": [{"id": 793934832, "item_id": 49620927, "url": "https://news.ycombinator.com/item?id=49620927", "by": "vvern", "date": "2026-09-09", "stance": "recommend", "quote": "I used to be a hater, but it\u2019s kinda good now", "excerpt": "I used to be a hater, but it\u2019s kinda good now", "context": "", "use_case": "", "switched_from": "", "story": "Cognition (Devin) raises $2B at $48B valuation"}, {"id": 794424784, "item_id": 49651549, "url": "https://news.ycombinator.com/item?id=49651549", "by": "deet", "date": "2026-09-10", "stance": "recommend", "quote": "at this point is one of the best cloud-hosted, team-managed coding agents", "excerpt": "As others have mentioned, it's really matured a lot and at this point is one of the best cloud-hosted, team-managed coding agents, when factoring overall UX, testing and QA lifecycle via its sandboxes, and its ability to be controlled with an API. We use it quite heavily. It's coming from a different starting place than Claude Code or Codex are as individually controlled single-developer tools. Devin has been more pe", "context": "team using it heavily", "use_case": "team-managed autonomous coding agent", "switched_from": "", "story": "Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra"}, {"id": 794374720, "item_id": 49648420, "url": "https://news.ycombinator.com/item?id=49648420", "by": "chris_st", "date": "2026-09-10", "stance": "recommend", "quote": "I use it, and have been happy with it for the most part.", "excerpt": "I use it, and have been happy with it for the most part. Like sibling, I use it for GPT-5.6-Sol and Opus work, and use their free models (GLM 5.2 for the past few months, trying SWE-2 now).", "context": "", "use_case": "accessing models like GPT and Opus", "switched_from": "", "story": "Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra"}]}, {"label": "Beats Copilot", "summary": "An enterprise user calls it a great product compared to Copilot.", "quotes": [{"id": 794391312, "item_id": 49649457, "url": "https://news.ycombinator.com/item?id=49649457", "by": "hightrix", "date": "2026-09-10", "stance": "recommend", "quote": "It's a great product compared to Copilot.", "excerpt": "My company uses it. We have a bunch of seats in an enterprise plan and have been using it for 9 months or so. It's a great product compared to Copilot. It is also the first AI tool I used heavily outside of creating random images or one off questions. I'm now using all three, Devin, Claude, Codex. I'm finding Claude and Codex to be much better. One of my biggest gripes is that the web client and desktop client for De", "context": "company with enterprise plan seats", "use_case": "AI coding at company", "switched_from": "", "story": "Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra"}]}], "caveats": [{"label": "Weak results", "summary": "Some found it poor even after re-evaluation, or only useful for trivial tasks.", "quotes": [{"id": 794413664, "item_id": 49650854, "url": "https://news.ycombinator.com/item?id=49650854", "by": "SaltyBackendGuy", "date": "2026-09-10", "stance": "warn", "quote": "It sucked the second time also...", "excerpt": "> And yes, I tried again This made me laugh a bit. I was forced to do an evaluation of their shit product twice due to being backed by the same PE firm; \"take a look at it again, it's much better now\". It sucked the second time also...", "context": "forced to evaluate it twice due to PE firm backing", "use_case": "AI coding agent evaluation", "switched_from": "", "story": "Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra"}, {"id": 738255361, "item_id": 46140960, "url": "https://news.ycombinator.com/item?id=46140960", "by": "rr808", "date": "2025-12-03", "stance": "warn", "quote": "most people can't even get the tools to do anything useful more than trivial things", "excerpt": "We have these weekly rah rah AI meetings where we swap tips on what we've achieved with copilot and devin. Mostly crickets but everyone is talking with lots of enthusiasm. Its starting to get silly though now, most people can't even get the tools to do anything useful more than trivial things we used to see on stack overflow.", "context": "employee at a company with weekly AI meetings", "use_case": "AI coding at work", "switched_from": "", "story": "Everyone in Seattle hates AI"}]}, {"label": "Behind competitors", "summary": "Some users find Claude Code and Codex better, and question its hands-off PR approach.", "quotes": [{"id": 794391312, "item_id": 49649457, "url": "https://news.ycombinator.com/item?id=49649457", "by": "hightrix", "date": "2026-09-10", "stance": "recommend", "quote": "It's a great product compared to Copilot.", "excerpt": "My company uses it. We have a bunch of seats in an enterprise plan and have been using it for 9 months or so. It's a great product compared to Copilot. It is also the first AI tool I used heavily outside of creating random images or one off questions. I'm now using all three, Devin, Claude, Codex. I'm finding Claude and Codex to be much better. One of my biggest gripes is that the web client and desktop client for De", "context": "company with enterprise plan seats", "use_case": "AI coding at company", "switched_from": "", "story": "Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra"}, {"id": 776941568, "item_id": 48558848, "url": "https://news.ycombinator.com/item?id=48558848", "by": "jatora", "date": "2026-06-16", "stance": "warn", "quote": "I cant believe people still use Devin.", "excerpt": "I cant believe people still use Devin. This is just filling the niche of vibe PR'ing? Not even vibe coding? Another level removed from the codebase than even claude code/codex are?", "context": "", "use_case": "", "switched_from": "", "story": "SpaceX to buy Cursor AI coding agent operator Anysphere for $60B"}]}, {"label": "Inconsistent clients", "summary": "Web and desktop clients use different harnesses, so response quality varies greatly.", "quotes": [{"id": 794391312, "item_id": 49649457, "url": "https://news.ycombinator.com/item?id=49649457", "by": "hightrix", "date": "2026-09-10", "stance": "recommend", "quote": "It's a great product compared to Copilot.", "excerpt": "My company uses it. We have a bunch of seats in an enterprise plan and have been using it for 9 months or so. It's a great product compared to Copilot. It is also the first AI tool I used heavily outside of creating random images or one off questions. I'm now using all three, Devin, Claude, Codex. I'm finding Claude and Codex to be much better. One of my biggest gripes is that the web client and desktop client for De", "context": "company with enterprise plan seats", "use_case": "AI coding at company", "switched_from": "", "story": "Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra"}]}, {"label": "Black box", "summary": "Criticized as too opaque and prone to hallucinating on complex legacy code.", "quotes": [{"id": 751799202, "item_id": 46987450, "url": "https://news.ycombinator.com/item?id=46987450", "by": "Priset-AI", "date": "2026-02-12", "stance": "warn", "quote": "I found existing AI coding tools (like Claude Code or Devin) too \"black box.\"", "excerpt": "Hi HN, I\u2019m Toni, a developer based in Brisbane, Australia. We just released the Priset extension for JetBrains/IntelliJ. The Problem: Like many of you, I found existing AI coding tools (like Claude Code or Devin) too \"black box.\" They often hallucinate on complex legacy code, and for enterprise work, sending code to be trained on by LLMs is a non-starter. We built Priset to be a 'Glass Box'\u2014it keeps the human in the ", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Show HN: Priset\u2013AI coding agent 4 IntelliJ,VSCode tht doesn't train on your code"}]}, {"label": "Vendor lock-in", "summary": "One commenter warns against depending on a single provider like Devin.", "quotes": [{"id": 801868129, "item_id": 50116758, "url": "https://news.ycombinator.com/item?id=50116758", "by": "tarasyarema", "date": "2026-10-16", "stance": "compare", "quote": "you should be free and not tied to a single provider (e.g. Devin)", "excerpt": "Hi there, Disclaimer: the blog should be glanced high-level as human, read the below instead. You can point your agents to it tho. We've been working on agent-swarm (https://github.com/desplega-ai/agent-swarm) as an oss MIT project for a while. One of the core principles on why we believe in is that when you adopt AI internally in your company, you should be free and not tied to a single provider (e.g. Devin). Yester", "context": "", "use_case": "on-call agent", "switched_from": "", "story": "Open Oncall: Open-Source Alternative to Devin Oncall"}]}], "who": [{"by": "deet", "context": "team using it heavily", "url": "https://news.ycombinator.com/item?id=49651549", "date": "2026-09-10"}, {"by": "hightrix", "context": "company with enterprise plan seats", "url": "https://news.ycombinator.com/item?id=49649457", "date": "2026-09-10"}, {"by": "PolCPP", "context": "grandfathered paying user", "url": "https://news.ycombinator.com/item?id=49647937", "date": "2026-09-10"}], "tags": ["for teams", "for enterprises", "API-controllable", "cloud-hosted"], "use_cases": ["on-call agent", "porting cado-nfs to gpus", "pr code review", "team-managed autonomous coding agent", "ai coding agent evaluation", "ai coding at company"], "compared": [{"slug": "codex", "name": "Codex", "co": 5, "switched_from": 0, "switched_to": 0}, {"slug": "claude-code", "name": "Claude Code", "co": 3, "switched_from": 0, "switched_to": 0}, {"slug": "cursor", "name": "Cursor", "co": 2, "switched_from": 0, "switched_to": 0}, {"slug": "copilot", "name": "Copilot", "co": 2, "switched_from": 0, "switched_to": 0}, {"slug": "github-copilot", "name": "GitHub Copilot", "co": 2, "switched_from": 0, "switched_to": 0}], "makers": [], "threads": [{"url": "https://news.ycombinator.com/item?id=49645443", "title": "Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra", "mentions": 5}, {"url": "https://news.ycombinator.com/item?id=50116757", "title": "Open Oncall: Open-Source Alternative to Devin Oncall", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49771966", "title": "RSA-896", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49703003", "title": "GPT-5.6 Luna vs. GPT-6 Astra: Is a $1.20 Model Good Enough for Code Review?", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49613312", "title": "Cognition (Devin) raises $2B at $48B valuation", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=48553224", "title": "SpaceX to buy Cursor AI coding agent operator Anysphere for $60B", "mentions": 1}], "sentiment_by_month": {"2025-11": {}, "2025-12": {"warn": 1}, "2026-01": {"neutral": 1, "recommend": 1}, "2026-02": {"warn": 1}, "2026-03": {}, "2026-04": {}, "2026-05": {}, "2026-06": {"warn": 1}, "2026-07": {}, "2026-08": {}, "2026-09": {"recommend": 4, "neutral": 2, "warn": 1, "compare": 1}, "2026-10": {"compare": 1}}, "quotes": [{"id": 801868129, "item_id": 50116758, "url": "https://news.ycombinator.com/item?id=50116758", "by": "tarasyarema", "date": "2026-10-16", "stance": "compare", "quote": "you should be free and not tied to a single provider (e.g. Devin)", "excerpt": "Hi there, Disclaimer: the blog should be glanced high-level as human, read the below instead. You can point your agents to it tho. We've been working on agent-swarm (https://github.com/desplega-ai/agent-swarm) as an oss MIT project for a while. One of the core principles on why we believe in is that when you adopt AI internally in your company, you should be free and not tied to a single provider (e.g. Devin). Yester", "context": "", "use_case": "on-call agent", "switched_from": "", "story": "Open Oncall: Open-Source Alternative to Devin Oncall"}, {"id": 796400865, "item_id": 49775054, "url": "https://news.ycombinator.com/item?id=49775054", "by": "bertonvv", "date": "2026-09-20", "stance": "neutral", "quote": "Devin (their AI agent) ported CADO-NFS to run on GPUs", "excerpt": "It seems that Eric Lu at Cognition AI used the exact same strategy on fewer GPUs to factor RSA-260 a couple weeks ago: https://cognition.com/blog/factoring-rsa-260 Devin (their AI agent) ported CADO-NFS to run on GPUs, similarly without any claimed algorithmic factoring improvements, they just let it run for 13 GPU-years. I recommend reading their article since it's much more thorough on details.", "context": "", "use_case": "porting CADO-NFS to GPUs", "switched_from": "", "story": "RSA-896"}, {"id": 795263490, "item_id": 49703968, "url": "https://news.ycombinator.com/item?id=49703968", "by": "dlahoda", "date": "2026-09-14", "stance": "compare", "quote": "finds bugs missed by coderabitai, devin-integration-bot, codex and copilot", "excerpt": "I would and do. works perfect https://github.com/dzmitry-lahoda/dz/tree/main/agents/skills/code-review burns half of day sub of astra for 200 USD. runs 1 hour on our repo. finds bugs missed by coderabitai, devin-integration-bot, codex and copilot (I ask them first until nothing found, yet my orchestration finds more). I do not hardcode our application in orchestration, but run subagent for applicationdomainproduct de", "context": "", "use_case": "PR code review", "switched_from": "", "story": "GPT-5.6 Luna vs. GPT-6 Astra: Is a $1.20 Model Good Enough for Code Review?"}, {"id": 794424784, "item_id": 49651549, "url": "https://news.ycombinator.com/item?id=49651549", "by": "deet", "date": "2026-09-10", "stance": "recommend", "quote": "at this point is one of the best cloud-hosted, team-managed coding agents", "excerpt": "As others have mentioned, it's really matured a lot and at this point is one of the best cloud-hosted, team-managed coding agents, when factoring overall UX, testing and QA lifecycle via its sandboxes, and its ability to be controlled with an API. We use it quite heavily. It's coming from a different starting place than Claude Code or Codex are as individually controlled single-developer tools. Devin has been more pe", "context": "team using it heavily", "use_case": "team-managed autonomous coding agent", "switched_from": "", "story": "Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra"}, {"id": 794413664, "item_id": 49650854, "url": "https://news.ycombinator.com/item?id=49650854", "by": "SaltyBackendGuy", "date": "2026-09-10", "stance": "warn", "quote": "It sucked the second time also...", "excerpt": "> And yes, I tried again This made me laugh a bit. I was forced to do an evaluation of their shit product twice due to being backed by the same PE firm; \"take a look at it again, it's much better now\". It sucked the second time also...", "context": "forced to evaluate it twice due to PE firm backing", "use_case": "AI coding agent evaluation", "switched_from": "", "story": "Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra"}, {"id": 794391312, "item_id": 49649457, "url": "https://news.ycombinator.com/item?id=49649457", "by": "hightrix", "date": "2026-09-10", "stance": "recommend", "quote": "It's a great product compared to Copilot.", "excerpt": "My company uses it. We have a bunch of seats in an enterprise plan and have been using it for 9 months or so. It's a great product compared to Copilot. It is also the first AI tool I used heavily outside of creating random images or one off questions. I'm now using all three, Devin, Claude, Codex. I'm finding Claude and Codex to be much better. One of my biggest gripes is that the web client and desktop client for De", "context": "company with enterprise plan seats", "use_case": "AI coding at company", "switched_from": "", "story": "Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra"}, {"id": 794374720, "item_id": 49648420, "url": "https://news.ycombinator.com/item?id=49648420", "by": "chris_st", "date": "2026-09-10", "stance": "recommend", "quote": "I use it, and have been happy with it for the most part.", "excerpt": "I use it, and have been happy with it for the most part. Like sibling, I use it for GPT-5.6-Sol and Opus work, and use their free models (GLM 5.2 for the past few months, trying SWE-2 now).", "context": "", "use_case": "accessing models like GPT and Opus", "switched_from": "", "story": "Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra"}, {"id": 794366992, "item_id": 49647937, "url": "https://news.ycombinator.com/item?id=49647937", "by": "PolCPP", "date": "2026-09-10", "stance": "neutral", "quote": "I pay for it (mostly because they grandfathered me from the old prices)!", "excerpt": "I pay for it (mostly because they grandfathered me from the old prices)! I used to use windsurf as my main editor until they changed their pricing model. Now i use it just to burn my weekly tokens on fable/astra if i remember to that on a task and that's it.", "context": "grandfathered paying user", "use_case": "burning weekly tokens on models", "switched_from": "", "story": "Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra"}, {"id": 793934832, "item_id": 49620927, "url": "https://news.ycombinator.com/item?id=49620927", "by": "vvern", "date": "2026-09-09", "stance": "recommend", "quote": "I used to be a hater, but it\u2019s kinda good now", "excerpt": "I used to be a hater, but it\u2019s kinda good now", "context": "", "use_case": "", "switched_from": "", "story": "Cognition (Devin) raises $2B at $48B valuation"}, {"id": 776941568, "item_id": 48558848, "url": "https://news.ycombinator.com/item?id=48558848", "by": "jatora", "date": "2026-06-16", "stance": "warn", "quote": "I cant believe people still use Devin.", "excerpt": "I cant believe people still use Devin. This is just filling the niche of vibe PR'ing? Not even vibe coding? Another level removed from the codebase than even claude code/codex are?", "context": "", "use_case": "", "switched_from": "", "story": "SpaceX to buy Cursor AI coding agent operator Anysphere for $60B"}, {"id": 751799202, "item_id": 46987450, "url": "https://news.ycombinator.com/item?id=46987450", "by": "Priset-AI", "date": "2026-02-12", "stance": "warn", "quote": "I found existing AI coding tools (like Claude Code or Devin) too \"black box.\"", "excerpt": "Hi HN, I\u2019m Toni, a developer based in Brisbane, Australia. We just released the Priset extension for JetBrains/IntelliJ. The Problem: Like many of you, I found existing AI coding tools (like Claude Code or Devin) too \"black box.\" They often hallucinate on complex legacy code, and for enterprise work, sending code to be trained on by LLMs is a non-starter. We built Priset to be a 'Glass Box'\u2014it keeps the human in the ", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Show HN: Priset\u2013AI coding agent 4 IntelliJ,VSCode tht doesn't train on your code"}, {"id": 745803440, "item_id": 46612715, "url": "https://news.ycombinator.com/item?id=46612715", "by": "cloudking", "date": "2026-01-13", "stance": "recommend", "quote": "We use https://devin.ai for this and it works very well.", "excerpt": "We use https://devin.ai for this and it works very well. Devin has it's own virtual environment, IDE, terminal and browser. You can configure it to run your application and connect to whatever it needs. Devin can modify the app, test changes in the browser and send you a screen recording of the working feature with a PR.", "context": "", "use_case": "background coding agent", "switched_from": "", "story": "Why We Built Our Own Background Agent"}, {"id": 743903793, "item_id": 46493987, "url": "https://news.ycombinator.com/item?id=46493987", "by": "thesandlord", "date": "2026-01-04", "stance": "neutral", "quote": "drafting a prompt for an AI agent (Cursor/Devin/Codex/etc) and letting it run", "excerpt": "I know everyone here is telling you not to hyper optimize, but I've found drafting a prompt for an AI agent (Cursor/Devin/Codex/etc) and letting it run and then using the remaining time to do a quick body weight set (pushups, squats)", "context": "", "use_case": "running AI agent during short gaps", "switched_from": "", "story": "Ask HN: How do you use 5\u201310 minute gaps productively?"}, {"id": 738255361, "item_id": 46140960, "url": "https://news.ycombinator.com/item?id=46140960", "by": "rr808", "date": "2025-12-03", "stance": "warn", "quote": "most people can't even get the tools to do anything useful more than trivial things", "excerpt": "We have these weekly rah rah AI meetings where we swap tips on what we've achieved with copilot and devin. Mostly crickets but everyone is talking with lots of enthusiasm. Its starting to get silly though now, most people can't even get the tools to do anything useful more than trivial things we used to see on stack overflow.", "context": "employee at a company with weekly AI meetings", "use_case": "AI coding at work", "switched_from": "", "story": "Everyone in Seattle hates AI"}]}