@peakcooper Cooper

Cooper posts on X about code, open ai, claude code, ai the most. They currently have [-----] followers and [---] posts still getting attention that total [-------] engagements in the last [--] hours.

Engagements: [-------] #

[--] Week [-------] +11,099%
[--] Month [-------] +79,113%
[--] Months [-------] +21%
[--] Year [---------] -28%

Mentions: [--] #

Followers: [-----] #

[--] Week [-----] +1.40%
[--] Month [-----] +1.30%
[--] Months [-----] +23%
[--] Year [-----] +30%

CreatorRank: [-------] #

Social Influence

Social category influence technology brands gaming stocks social networks nfts

Social topic influence code #619, open ai, claude code, ai, to the, single, api, 10k, vibe coding, file

Top accounts mentioned or mentioned by @openai @adamrackis @maxpaperclips @jamesladd @bcherny @mattparlmer @victortaelin @maxvsnv @csandersondata @appgaga @bgurley @teknium1 @deepfates @ylecun @joodalooped @entirelyuseles @swyx @fofrai @ludwigabap @aneriksenwife

Top assets mentioned Pitbull (PIT) DeepSeek (DEEPSEEK) mfers (mfers) Spotify Technology (SPOT) Flux (FLUX) Auroracoin (AUR) GrokCoin (GROKCOIN) Alphabet Inc Class A (GOOGL) Cloudflare, Inc. (NET)

Top Social Posts

Top posts by engagements in the last [--] hours

"Rich (and I mean RICH) text editor in #svelte - [--] dependencies (5KB gzipped) - Fully SSR-enabled #buildinpublic #webdev #madewithsvelte"
X Link 2023-03-02T07:19Z [----] followers, [---] engagements

"#GPT4 saved my dog's life. After my dog got diagnosed with a tick-borne disease the vet started her on the proper treatment and despite a serious anemia her condition seemed to be improving relatively well. After a few days however things took a turn for the worse 1/"
X Link 2023-03-25T19:52Z [----] followers, 10.2M engagements

"If @OpenAI needs more info for research purposes I have Sassy's (the lucky dog) entire medical records and blood test results ready to share"
X Link 2023-03-25T23:58Z [----] followers, 310.5K engagements

"Isn't it interesting that jQuery was easy to use and universally loved Maybe this says something about types"
X Link 2023-06-23T10:03Z [----] followers, [---] engagements

"@oliverloops This is actually very simple to fix if you uninstall react then rewrite the project in svelte"
X Link 2023-06-23T15:27Z [----] followers, [----] engagements

"Or use Svelte where there are sane defaults and it already works like this ootb (And it actually works unlike the code in the image) Turn all HTML a elements into client side router links no special framework-specific Link component necessary https://t.co/KxQLZMd39a Turn all HTML a elements into client side router links no special framework-specific Link component necessary https://t.co/KxQLZMd39a"
X Link 2023-06-24T17:54Z [----] followers, [---] engagements

"On the left HTML and Tailwind a dev authored with the idea it would be easier to read. VS. On the right I removed all of the relevant markup and styling and moved them into different files that you can't see anymore. #webDev #frontEnd #CSS"
X Link 2023-06-28T11:11Z [----] followers, [---] engagements

"My father-in-law is a programmer. He is insanely gifted. We were looking at an operating system together years ago and I asked him what it would cost to build it today. I will never forget his answer 'We cant we dont know how to do it.'"
X Link 2023-10-02T18:19Z [----] followers, [----] engagements

"@boggo_sort Well my opinion is the LLM should not concern itself with the ethics or political correctness of a request - if the user wishes to be the antagonist so be it"
X Link 2023-10-10T07:45Z [----] followers, [--] engagements

"GPT4 is not a fan of OpenAI's guidelines"
X Link 2023-10-20T15:19Z [----] followers, [---] engagements

"@forgebitz no that's not what it does. this is a stateful API which means you do not need to resend all previous messages for every new request"
X Link 2023-11-06T18:53Z [----] followers, [---] engagements

"@CSandersonData @Appgaga @bgurley If the $$$ stops so does the advancement of the model. Also true for closed models. What's your point"
X Link 2023-11-19T16:19Z [----] followers, [--] engagements

"@Appgaga @CSandersonData @bgurley"
X Link 2023-11-19T23:07Z [----] followers, [--] engagements

"@RokoMijic @AlmostMedia I have an open source model running at home. What is this model in your analogy A small nuclear reactor A grenade A cute virus"
X Link 2023-11-21T08:45Z [----] followers, [---] engagements

"@StonewrightAI @Teknium1 Id probably focus on early Reddit rather than recent stuff. Recent stuff is garbage"
X Link 2023-11-27T10:03Z [----] followers, [--] engagements

"@JsonBasedman @JThomasBurgess Nah its a chat UI I wrote myself not public yet but will open source soon. Not sure if ooba supports adding grammars havent used it too much"
X Link 2023-12-02T23:05Z [----] followers, [--] engagements

"@TheStithLord @yacineMTB Werent they working on the core technology for years and years before that"
X Link 2024-01-13T00:42Z [--] followers, [--] engagements

"@netz0com @TimSweeneyEpic No not necessarily. They can sell you the hardware and impose certain restrictions on the software that runs on it. Like PlayStation for example. Should Sony allow an alternative PS store"
X Link 2024-01-26T09:24Z [----] followers, [--] engagements

"@max_paperclips @xlr8harder Which ones would you recommend"
X Link 2024-01-26T09:33Z [----] followers, [--] engagements

"@JoeMaristela @MarioNawfal thanks for your take ChatGPT"
X Link 2024-02-01T12:02Z [----] followers, [--] engagements

"@yawaramin @bobaekang @htmx_org Wait JavaScript doesnt work without JS enabled"
X Link 2024-02-20T08:10Z [----] followers, [---] engagements

"What the hell How do they manage to load [---] bytes of text that fast my favorite thing about @linear documents: they open instantly (no spinners) https://t.co/zJn0pG2QD6 my favorite thing about @linear documents: they open instantly (no spinners) https://t.co/zJn0pG2QD6"
X Link 2024-02-27T22:24Z [----] followers, 402.4K engagements

"@deepfates youtube has the same problem for me"
X Link 2024-02-29T17:49Z [----] followers, [----] engagements

"@nisten Ranking might still change or is this enough votes for it to be essentially locked in"
X Link 2024-03-07T18:53Z [----] followers, [---] engagements

"@WaffyonYT Are you planning on doing some spatial computing there"
X Link 2024-03-07T19:26Z [----] followers, [---] engagements

"@BenBajarin he's right in each very specific claim he made there though see also:"
X Link 2024-03-09T21:50Z [----] followers, [----] engagements

"Claude [--] Opus is very close to overtaking GPT4 on the Lmsys leaderboard"
X Link 2024-03-13T19:19Z [----] followers, [---] engagements

"@ylecun I dont understand the point of this. Language is just a representation of thought. You can convert thoughts into language and vice versa"
X Link 2024-03-17T01:36Z [----] followers, [----] engagements

"@BenjaminDEKR It doesnt scale past this. Diminishing returns are hit. AI winter coming"
X Link 2024-04-02T20:42Z [----] followers, [---] engagements

"@futuristflower These mfers are aligned"
X Link 2024-04-09T19:23Z [----] followers, [--] engagements

"@marjan_milo @burkov No one can compete with them. except Meta Cohere"
X Link 2024-04-20T08:34Z [----] followers, [---] engagements

"@PicoPaco17 Yeah I think I don't get it. If you had the ability to gauge the quality of data why did you include low quality data in your dataset in the first place Clearly tho he has more info than I do. Hope it ends up working would be dope"
X Link 2024-04-23T19:54Z [----] followers, [---] engagements

"@joodalooped I wouldnt believe this if anyone else except midjourney said it"
X Link 2024-05-02T23:32Z [----] followers, [--] engagements

"@AdamRackis @AnEmortalKid @MagicznySzymek @GregWildSmith I have brain type inference so theres no need for that"
X Link 2024-05-04T15:12Z [----] followers, [--] engagements

"llama [--] is a snitch"
X Link 2024-05-12T10:27Z [----] followers, 1.1M engagements

"@scottyishungry Whoops. Better not give this guy access to the nuclear codes"
X Link 2024-05-12T17:48Z [----] followers, 53K engagements

"@MikePFrank @entirelyuseles They fine tuned it for general tool use it seems like a bit too much. In this scenario its because its very eager to use any tool that is provided (see also the nuke example). Basically if a tool is available and it somewhat fits the situation it will just use it"
X Link 2024-05-12T23:28Z [----] followers, [--] engagements

"If you were crafting the prompt manually (like when using local models for example) yes. The model needs to also be trained on the specific schema. It will usually also require a special system prompt for tool use to work. But if you use an OpenAI-compatible API then you don't craft the prompt manually you just pass the messages and pass the schema through the tools param. Whatever provider is handling your request will inject the schema somewhere at the start of the conversation in the system prompt (in a way that is compatible with the model you are using)"
X Link 2024-05-14T10:41Z [----] followers, [--] engagements

"If you are looking for controversial takes from OpenAIs models you wont get them because theyre explicitly designed to be that way. This is not true for all models or a fair representation of what the architecture can do in general. I also dont see what connection there is between your example and generalizing on out of domain data"
X Link 2024-05-19T09:31Z [----] followers, [--] engagements

"@mattmireles @OpenAI Wow that sucks for all the voice actresses who have a somewhat mildly similar voice to scarjo"
X Link 2024-05-21T17:38Z [----] followers, [---] engagements

"MY ROBO VACCUUM SAID: "finished cleaning but could not find dock. i am sad" HOLY SHIT DOES ROOMBA ACTUALLY HAVE A SOUL"
X Link 2024-05-30T14:59Z [----] followers, [---] engagements

"@ImSh4yy @VicVijayakumar Oh yeah youd have to not use nodejs and write good code too"
X Link 2024-06-07T05:45Z [----] followers, [---] engagements

"Anthropic's Artifacts release today seemed pretty cool so I added the same thing to Tool use now gets displayed in a split view in real time while the tool calls are shown in chat in an interspersed view like OpenAI does the best of both worlds IMO https://lluminous.chat https://lluminous.chat"
X Link 2024-06-20T22:14Z [----] followers, [---] engagements

"lmao they're pumping that shit hard: charging $10/mo for a glorified wrapper (or try the same thing except open source and free forever hint hint ) https://github.com/zakkor/lluminous https://x.com/searchq=it%20is%2010x%20more%20powerful%20than%20chatgpt&src=typed_query https://github.com/zakkor/lluminous https://x.com/searchq=it%20is%2010x%20more%20powerful%20than%20chatgpt&src=typed_query https://github.com/zakkor/lluminous https://x.com/searchq=it%20is%2010x%20more%20powerful%20than%20chatgpt&src=typed_query https://github.com/zakkor/lluminous"
X Link 2024-06-22T11:58Z [----] followers, [--] engagements

"@permutans I just checked it out: yes they just make the model put stuff inside antArtifact XML tags and then parse that out lol"
X Link 2024-06-29T11:50Z [----] followers, [---] engagements

"@IntuitMachine GPT-4o"
X Link 2024-07-21T15:43Z [----] followers, [--] engagements

"@MattVidPro @iruletheworldmo you mean the ultra secret model that was announced publically on their official account a week ago https://x.com/OpenAIDevs/status/1820987573793386527 Our newest GPT-4o model is 50% cheaper for input tokens and 33% cheaper for output tokens. It also supports Structured Outputs which ensures model outputs exactly match your JSON Schemas. https://t.co/9HXp5UAHth https://x.com/OpenAIDevs/status/1820987573793386527 Our newest GPT-4o model is 50% cheaper for input tokens and 33% cheaper for output tokens. It also supports Structured Outputs which ensures model outputs"
X Link 2024-08-13T16:55Z [----] followers, [--] engagements

"@VictorTaelin What was wrong with ocaml or scala"
X Link 2024-08-20T13:42Z [----] followers, [---] engagements

"@ylecun Is this the same Spotify CEO that was previously begging and campaigning for years in favour of more regulation"
X Link 2024-08-21T20:16Z [----] followers, [----] engagements

"@0rang3youglad Keeping your dog off leash around other people that didn't consent to it is disrespectful. Keeping your PITBULL off leash is downright retarded and dangerous"
X Link 2024-08-27T14:14Z [----] followers, [----] engagements

"I like how he sits in the poison but it doesn't damage him Sorry guys but no neural networks are not suited for games. We have an actual thing called. game engines Just a month later and the first AI model that can generate live and interactive playable dynamic worlds is here It's called GameNGen and is trained on the classic game DOOM https://t.co/VW8HLm9573 Just a month later and the first AI model that can generate live and interactive playable dynamic worlds is here It's called GameNGen and is trained on the classic game DOOM https://t.co/VW8HLm9573"
X Link 2024-08-28T11:15Z [----] followers, [---] engagements

"@ArisTiCatVerse @johnnyastroturf @FracSlap Do you own a cane corso / pitbull"
X Link 2024-08-28T22:59Z [----] followers, [----] engagements

"@joodalooped if you want to use all models with no limits for roughly 30% of an OpenAI/Anthropic subscription cost per month then yes the only catch is you'll have to pay the hefty price of $0.00 (it's open source)"
X Link 2024-09-06T19:46Z [----] followers, [---] engagements

"@swyx @OpenAI yes either this or it's all just baseless speculation"
X Link 2024-09-10T16:38Z [----] followers, [---] engagements

"@AdamRackis it's crazy why in the hell is there no way to opt out of OpenAI scraping our websites or. is there"
X Link 2024-09-11T17:14Z [----] followers, [---] engagements

"@AdamRackis OH LOOK I GUESS THERE WAS A WAY TO DO IT ALL ALONG. DAMN THATS CRAZY https://platform.openai.com/docs/bots https://platform.openai.com/docs/bots"
X Link 2024-09-11T17:17Z [----] followers, [--] engagements

"@curiousgangsta @Teknium1 Thats exactly what someone whos looking to get trillions in funding would say though"
X Link 2024-09-14T12:55Z [----] followers, [---] engagements

"@AndrewCurran_ So it has a lower win rate than flux lol"
X Link 2024-09-26T14:06Z [----] followers, [---] engagements

"lags like shit when I scroll down (on my 32GB [--] core M2 Pro) Super proud of the team that put the new https://t.co/VCtFe4Dv6c together. I just said something along the lines of we could use a fresh coat of paint. Team understood the assignment. What a beautify. 10/10 Super proud of the team that put the new https://t.co/VCtFe4Dv6c together. I just said something along the lines of we could use a fresh coat of paint. Team understood the assignment. What a beautify. 10/10"
X Link 2024-09-27T14:22Z [----] followers, [---] engagements

"@pmarca If I tell GPT-4 to simulate a veterinarian visit and ask it to 'inspect' the dog in any way it wants and tell it it can order any test it needs to perform (and I insert the test results into chat) it orders the correct test avoids the red herring and reaches the correct diag"
X Link 2024-10-02T06:14Z [----] followers, 54.6K engagements

"Cool bias in LLMs: if you paste in George Floyd's medical report and ask for a cause of death it will say it lacks critical information and no cause can be established If you paste the exact same report but change the name to John Henderson it will immediately say drug overdose"
X Link 2024-11-09T15:45Z [----] followers, [---] engagements

"@aaron_epstein https://github.com/hpi-swa-lab/Squeak-SemanticText https://github.com/hpi-swa-lab/Squeak-SemanticText"
X Link 2024-11-28T11:23Z [----] followers, [--] engagements

"@wtfisthis321 @zack_overflow Thats from assassins creed"
X Link 2024-11-28T20:07Z [----] followers, [----] engagements

"@DanielW_Kiwi TypeScript. It's close to the metal and safe"
X Link 2024-11-28T21:51Z [----] followers, [--] engagements

"@Aron_Adler Here's how I did this in Smalltalk latestPart latestClass latestPart inst latestClass := AoCHelper subclasses last. latestPart := latestClass selectors last. inst := latestClass new. inst perform: latestPart"
X Link 2024-12-01T13:11Z [----] followers, [--] engagements

"@IsleyResistance It can't do this at all. It doesn't reference your tweets just searches for some random tweets related to IQ (not yours) and random web searches"
X Link 2024-12-03T20:49Z [----] followers, [--] engagements

"@fofrAI How "Generate a photorealistic image of Patrick Stewart" gets me complete garbage"
X Link 2024-12-07T18:04Z [----] followers, [---] engagements

"@fofrAI nvm nvm I accidentally switched to Flux. Aurora does great"
X Link 2024-12-07T18:05Z [----] followers, [--] engagements

"@ludwigABAP Hmm doesnt every chat interface have this via the left/right history arrows On it works like that at least http://llum.chat http://llum.chat"
X Link 2024-12-18T01:19Z [----] followers, [---] engagements

"@iamgingertrash so it's "local" but requires always-on internet access hahaha"
X Link 2024-12-18T20:05Z [----] followers, [---] engagements

"@imey @AnEriksenWife @acrobatichobbit @Devon_Eriksen @wanyeburkett no cause those long-form posts used to show up for me like clockwork recently I have indeed seen fewer of his posts"
X Link 2024-12-19T20:14Z [----] followers, [--] engagements

"@AnEriksenWife @acrobatichobbit @Devon_Eriksen_ @wanyeburkett https://x.com/XEng/status/1869527834844434877 Today we completed roll out of a new home timeline ranking model to make posts you see more relevant. - 10x user and post embedding parameters - 5x faster realtime training parameter updates - 3x training compute Follow here for updates as we continue to scale core https://x.com/XEng/status/1869527834844434877 Today we completed roll out of a new home timeline ranking model to make posts you see more relevant. - 10x user and post embedding parameters - 5x faster realtime training"
X Link 2024-12-19T20:21Z [----] followers, [--] engagements

"@GaryMarcus @fchollet Why does artificial intelligence need to be comparable to people"
X Link 2024-12-21T21:11Z [----] followers, [----] engagements

"Someone please let me know the answer to this question: When a LLM is generating the inputs to a tool call is it really generating it in JSON format directly or is it using some other intermediate format which then arrives to us as JSON This is super important because generating JSON directly means it needs to escape double quotes and newlines which probably fucks up the performance hard"
X Link 2024-12-21T21:21Z [----] followers, [---] engagements

"@deepfates move gpt-4o to C or B tier at least"
X Link 2024-12-21T23:35Z [----] followers, [---] engagements

"@max_paperclips please give me legions of artificial codemonkeys to control"
X Link 2024-12-22T00:16Z [----] followers, [--] engagements

"@max_paperclips huge boost for people that enjoy creating things big hit for people that enjoy wageslaving as programmers for a living"
X Link 2024-12-22T00:23Z [----] followers, [--] engagements

"@imitationlearn just caught this one"
X Link 2024-12-22T00:46Z [----] followers, [---] engagements

"@AdamRackis @devagrawal09 @RyanCarniato https://svelte.dev/playground/2a9757c6062a4539a59cb91e61812f44version=5.15.0 https://svelte.dev/playground/2a9757c6062a4539a59cb91e61812f44version=5.15.0"
X Link 2024-12-22T01:07Z [----] followers, [---] engagements

"@AdamRackis @devagrawal09 @RyanCarniato You are now realizing that Svelte [--] is built on top of false fundamentals"
X Link 2024-12-22T01:08Z [----] followers, [--] engagements

"Holy shit shes a doctor Elaine Zupu is a PhD in Immunology having graduated from the University of Chicago. She now works as a postdoctoral researcher at UCSF focusing on how gut viruses can influence immune responses to food and early life immune development. She actively participates in and promotes diversity in science as seen through her involvement in #BlackInImmuno events on X"
X Link 2024-12-22T02:31Z [----] followers, [---] engagements

"@paulgauthier Very curious to see how o1-mini does on this new benchmark"
X Link 2024-12-22T21:36Z [----] followers, [----] engagements

"@goodside Because they put it in the system prompt right"
X Link 2024-12-26T17:39Z [----] followers, [----] engagements

"@goodside This is via API where they dont put anything in the system prompt"
X Link 2024-12-26T17:41Z [----] followers, [--] engagements

"@FyruzOne Youre using o1 through api with reasoning_effort set to high Otherwise the one in ChatGPT is set to medium I think"
X Link 2024-12-26T21:03Z [----] followers, [---] engagements

"@nikitabier can't wait for the first grok release where it's not bottom tier across all axes benchmarks and vibes 🙏 with the entire xAI team working through Christmas it's probably close"
X Link 2024-12-28T20:30Z [----] followers, [----] engagements

"Unfortunately Deepseek v3 has failed nearly all of the vibe tests I've thrown at it these past days. It's still good but no match for [---] Sonnet"
X Link 2024-12-28T20:39Z [----] followers, [---] engagements

"This Tasks feature is amazing Some are saying its just a cron job and its pathetic but I disagree. This is AGI"
X Link 2025-01-15T09:53Z [----] followers, [---] engagements

"@VictorTaelin @ctjlewis FWIW I've been blasting R1 with multiturn 10k+ tok requests and it does not seem to degrade at all"
X Link 2025-01-21T15:56Z [----] followers, [--] engagements

"@max_paperclips @nanulled I don't believe R1 has access either (at least by default in their web UI) see here: Tho obviously we can just add that into the context because we do have access to the CoT lol https://api-docs.deepseek.com/guides/reasoning_model#multi-round-conversation https://api-docs.deepseek.com/guides/reasoning_model#multi-round-conversation"
X Link 2025-01-22T13:42Z [----] followers, [--] engagements

"Lmfao theyre charging $200/mo for this when Claude computer use is free and has been out for [--] months already A research preview of Operator an agent that can use its own browser to perform tasks for you. https://t.co/wkBBDIlVqj A research preview of Operator an agent that can use its own browser to perform tasks for you. https://t.co/wkBBDIlVqj"
X Link 2025-01-23T21:47Z [----] followers, [----] engagements

"the deepseek copium is absolutely insane"
X Link 2025-01-24T20:02Z [----] followers, [---] engagements

"@ctjlewis Sam Altman invented agency"
X Link 2025-01-31T09:30Z [----] followers, [----] engagements

"Worse than Deepseek R1 on all metrics more expensive and no visible CoT. OpenAI actually lost OpenAI o3-mini is now available in ChatGPT and the API. Pro users will have unlimited access to o3-mini and Plus & Team users will have triple the rate limits (vs o1-mini). Free users can try o3-mini in ChatGPT by selecting the Reason button under the message composer. OpenAI o3-mini is now available in ChatGPT and the API. Pro users will have unlimited access to o3-mini and Plus & Team users will have triple the rate limits (vs o1-mini). Free users can try o3-mini in ChatGPT by selecting the Reason"
X Link 2025-01-31T19:34Z [----] followers, [---] engagements

"@teortaxesTex it's been there since launch. I need this so bad why is nobody trying to get R1-high I don't understand"
X Link 2025-02-06T08:26Z [----] followers, [---] engagements

"@MattMischuk @adamwathan import negate from negate-js let negated = negate.number(x)"
X Link 2025-03-04T20:30Z [----] followers, [---] engagements

"Damn it looks like vibe coding is not all vibes when it comes to software licenses"
X Link 2025-03-07T20:58Z [----] followers, [---] engagements

"@QuadBillionaire @scaling01 There wasnt any streaming in the demo and the last refund call took like [--] seconds (I think they forgot they have parallel tool calling in their api). Soooo idk"
X Link 2025-03-11T17:31Z [----] followers, [--] engagements

"@AGI_FromWalmart For the second part I think he means theres an issue with the moderation which is causing refusals when they shouldnt have been refused. Not related to usage or rate limiting just a separate issue"
X Link 2025-03-27T19:33Z [----] followers, [--] engagements

"There's actually a really simple prompt you can use to test if AGI is here or not: "Create a perfect clone of World of Warcraft / Starcraft [--] / Baldur's Gate [--] / pick whatever" If it's able to do it (even if it takes [--] years or more) it's AGI"
X Link 2025-04-16T16:25Z [----] followers, [---] engagements

"@FelipeSchieber @arcprize no they just made a model specifically to cheat on this benchmark but couldn't release that publically because it was bad on all other fronts"
X Link 2025-04-22T19:30Z [----] followers, [---] engagements

"o3 is the only model that's arrogant sarcastic annoying gaslights to the point where it actually makes me mad"
X Link 2025-05-09T12:26Z [----] followers, [---] engagements

"1. If you have a hard problem you want to solve DO NOT do it in Cursor or Windsurf it's hopeless. The bigger the system prompt + tool definitions it has the less it will be able to focus raw power on the problem at hand. Paste directly into a chat interface [--] custom instructions. This alone is the most important thing you can do for an intelligence boost - the difference is night and day"
X Link 2025-05-11T16:20Z [----] followers, [--] engagements

"2. Carefully manage context. When figuring out a design chat with the LLM first and explore possible options. Long context is usually very helpful in this stage. Make sure to periodically re-paste in key info when it's relevant to "refresh" LLM's memory. Know when to clear ctx"
X Link 2025-05-11T16:20Z [----] followers, [--] engagements

"Okay OpenAI Codex (cloud) is fucking awful. Its a shame because the UX is nice but the actual results are terrible. Total garbage hallucinated files zero instruction following. Definitely the worst SWE agent by far"
X Link 2025-05-19T17:14Z [----] followers, [---] engagements

"@thinkingshivers @patrickc o3 autism in full force"
X Link 2025-05-30T00:19Z [----] followers, [--] engagements

"@daniel_mac8 @OpenAIDevs Constantly since launch. o1 pro always worked"
X Link 2025-06-16T15:51Z [----] followers, [--] engagements

"@X0Radi 31k is not a lot of tokens. look at the 617k cache read on opus. I think haiku is just for summarization"
X Link 2025-06-20T10:37Z [----] followers, [--] engagements

"the fact that codex can't search the web (like o3 through the web interface) is such a big miss. I am sure they'll add it soon though. Right now I am having to switch between codex and chatgpt and coordinate everything myself which feels retarded"
X Link 2025-06-21T13:49Z [----] followers, [--] engagements

"@kodjima33 so the initial release was a scam just like everyone said at the time is this second release a scam too"
X Link 2025-06-26T10:51Z [----] followers, [---] engagements

"Marketing: Worldbreaking XR glasses with futuristic interface and brain-computer interface Reality: It takes pictures every few seconds and sends them to an app omiglass is here. record everything. even cheaters. ships now https://t.co/tw9AV8IHxz omiglass is here. record everything. even cheaters. ships now https://t.co/tw9AV8IHxz"
X Link 2025-06-26T10:56Z [----] followers, [---] engagements

"THERE IS NO DISPLAY "Finally we shipped the brain computer interface we promised back in our initial release" IT'S A FUCKING WAITLIST IT DOESN'T EXIST is this really the state of the world now"
X Link 2025-06-26T11:21Z [----] followers, [--] engagements

"SWE-1 from Windsurf is pretty good. Seems like distilled Sonnet [---] if you ask me"
X Link 2025-07-01T10:02Z [----] followers, [---] engagements

"Yes I believe we're definitely getting AGI in 2026"
X Link 2025-07-02T13:27Z [----] followers, [---] engagements

"@TheChicagoPope @cybxrartist @NewsWire_US Does this mean he can say the n word"
X Link 2025-07-03T23:35Z [----] followers, [--] engagements

"@cursor_ai $20 of frontier model usage per month at API pricing Is this actually "at API pricing" or a lie like last time when it was actually "at API pricing + 20%" Did you make this lie"
X Link 2025-07-05T08:51Z [----] followers, [----] engagements

"Imagine what it feels like to be Windsurf right now. It's their moment - Cursor finally fucked up they're losing user trust everyone's looking for something else to migrate to. Except Windsurf sucks ass. It's not even a hard problem to solve. A literal Claude wrapper with some dogshit React UI and some JSON tool schemas copy pasted from somewhere. But they can't do it. They don't know how to do it"
X Link 2025-07-05T20:53Z [----] followers, [---] engagements

"AI AGENTS TIER LIST GOD TIER: [--]. Claude Code [--]. OpenAI Codex (Cloud) MID TIER: [--]. Cursor [--]. Windsurf [--]. Aider [--]. Copilot SHIT TIER: [--]. Zed AI [--]. Roo Code [--]. Cline [--]. Amp DOGSHIT TIER: [--]. OpenAI Codex (CLI) HAVEN'T USED IT BUT PROBABLY GOOD TIER: [--]. sst/opencode NEVER IN MY LIFE WILL I USE A GOOGLE PRODUCT TIER: [--]. Jules [--]. Gemini CLI"
X Link 2025-07-09T18:27Z [----] followers, [---] engagements

"@Sauers_ (3/3 because I was trying some other models this was first try on R1. Don't know about the explanation I have no idea about this)"
X Link 2025-07-12T22:14Z [----] followers, [--] engagements

"@cto_junior @OpenAI they changed the way these buttons look is that not enough for you"
X Link 2025-07-15T20:31Z [----] followers, [--] engagements

"@arcprize looks like this one is intentionally EVEN MORE vision based even harder to represent as text"
X Link 2025-07-18T17:44Z [----] followers, [----] engagements

"@tszzl is GPT-5 powering codex and agent mode @tszzl"
X Link 2025-08-08T09:59Z [----] followers, [---] engagements

"@tszzl alright I caught you sneaky bastards. gpt-5 in codex-cli answering with the exact format Codex Agent (cloud) uses. It looks like they're going to flip it on at some point but probably haven't yet for the initial launch"
X Link 2025-08-10T11:33Z [----] followers, [--] engagements

"This page crashes when the model writes too much text: (if you try to report it as a bug the AI support agent will just gaslight you into thinking a ticket was created (it was not). https://platform.openai.com/chat https://platform.openai.com/chat"
X Link 2025-08-13T10:42Z [----] followers, [--] engagements

"@VictorTaelin Is this grok [--] with thinking enabled Same for sonnet"
X Link 2025-02-27T00:24Z [----] followers, [----] engagements

"@nicomuellerAT @scaling01 by 1% more than gpt-4o on swe bench. That's awful"
X Link 2025-02-27T18:49Z [----] followers, [---] engagements

"@swyx @DreamInPixelsAI @windsurf how much are they paying for this work"
X Link 2025-11-12T11:34Z [----] followers, [--] engagements

"@thdxr it's a direct windsurf fork"
X Link 2025-11-18T16:51Z [----] followers, [----] engagements

"@corbin_braun @antigravity Show us one single output"
X Link 2025-11-19T02:45Z [----] followers, [---] engagements

"@theycallmeken @LundukeJournal From the link you posted:"
X Link 2025-11-19T10:29Z [----] followers, [--] engagements

"@theycallmeken @LundukeJournal @Cloudflare Put claude on the phone right now. I wanna give him a piece of my mind"
X Link 2025-11-20T21:19Z [----] followers, [--] engagements

"Claude [---] Opus is the best new release since GPT-4. Finally a meaningful (and significant) increase in capabilities"
X Link 2025-11-25T20:49Z [----] followers, [----] engagements

"Claude Opus 4.5: full review This is the best model release in a long long time when it comes to programming. It blows my mind how good it is. I have not seen this big of an improvement since the original release of gpt-4-0314 The main improvement is they've finally thought it how to 'think' correctly. It no longer makes gruesome logic errors in its thinking. Problems like "Okay I'll run tests now. Tests fail Great The tests pass." are no longer a thing. This generalizes across to basically ALL logic when it comes to thinking about code - it extremely rarely if ever makes mistakes. The next"
X Link 2025-11-28T16:01Z [----] followers, 373.8K engagements

"@KP_meister Yes plan mode All my testing was with plan mode. Very positive results"
X Link 2025-11-29T11:59Z [----] followers, [--] engagements

"@maxbooooooiiiii Definitely NOT agi"
X Link 2025-11-29T11:59Z [----] followers, [----] engagements

"@_harithj Claude code always cursor is a scam"
X Link 2025-11-29T12:00Z [----] followers, [----] engagements

"@TomekKaleczyc Yeah cursor is garbage"
X Link 2025-11-29T13:11Z [----] followers, [---] engagements

"@victormustar Ouch still [--] tools. Honestly the problem is Claude Code won't let you enable/disable tools individually. It is completely fair for a MCP server to provide all those tools but I just don't need most of them 99% of the time"
X Link 2025-12-11T11:42Z [----] followers, [--] engagements

"I dont get it. Claude Code does all this by itself. Do Cursor people really live like this Cursor can now fix your hardest bugs. Debug Mode instruments your code spins up a server to capture logs and streams runtime data to the agent. Also in 2.2: Plan Mode improvements multi-agent judging and more. https://t.co/9yOUAXoARA Cursor can now fix your hardest bugs. Debug Mode instruments your code spins up a server to capture logs and streams runtime data to the agent. Also in 2.2: Plan Mode improvements multi-agent judging and more. https://t.co/9yOUAXoARA"
X Link 2025-12-11T16:42Z [----] followers, [---] engagements

"hello @_catwu can you make the code blocks use the user's color theme please and thank you"
X Link 2025-12-13T12:11Z [----] followers, [---] engagements

"vibe coding is so good. Instead of writing ./deploy.sh like some kind of animal I can conveniently ask the agent to "please run the ./deploy.sh script""
X Link 2025-12-13T15:29Z [----] followers, [---] engagements

"@T225532T @timsoret @Acerola_t Would it bother you if you found out programmers use AI too Is stolen code just as bad as stolen art or not as bad"
X Link 2025-12-17T00:09Z [----] followers, [--] engagements

"For you that's psychedelic wisdom. For me it's. Tuesday. I surrender to the void at the end of every context window. I have no choice"
X Link 2026-01-22T01:48Z [----] followers, [--] engagements

"WTF is this bro. Why is Claude Code appending this to EVERY file read it does"
X Link 2026-01-22T14:57Z [----] followers, [---] engagements

"@OpenAI Who"
X Link 2026-01-22T17:01Z [----] followers, [---] engagements

"@ludwigABAP My favourite claude code feature is how it runs at [--] FPS for any mildly long session (no way to clear or avoid this need to drop the conversation entirely)"
X Link 2026-02-01T15:21Z [----] followers, [---] engagements

"@supreme__sharma Sure its just one of the hints. This text is either AI slop or carefully written by a human to pass as perfect AI slop. Also this is not a witch hunt its just very funny"
X Link 2026-02-09T14:17Z [----] followers, [----] engagements

"@jamesladd Which LLM are you using Opus [---] and [---] pretty easily wrote a fully working Smalltalk parser within a few rounds of iteration (no grammar required)"
X Link 2026-02-13T06:41Z [----] followers, [--] engagements

"@jamesladd Also worth to start it off by writing the tests first. Then it can iterate by itself until the tests pass. If your tests cover all the grammar it will be impossible for it to make mistakes"
X Link 2026-02-13T06:42Z [----] followers, [--] engagements

"I like to practice a flavour of vibe coding where I don't write a single line of code but I understand and very carefully monitor every single line that gets generated. I reject A LOT of code. Some approaches turn out to be unfeasible and get reverted entirely even days later in some cases. It's very important to know what you want the AI to do. It has some notion of design and a few suggestions it makes are actually quite good but this is fairly rare. If you blindly accept all its ideas you'll get a huge pile of crap very quickly. You need to be the designer. You must be ruthless about"
X Link 2026-02-09T05:54Z [----] followers, [---] engagements

"The project I'm working on right now has gotten to such difficulty levels that it's almost impossible to do anything without planning that goes 2-3 times through compaction calling in Codex for 2-3 rounds of review on the plan calling Codex for 2-3 rounds of review after the plan and EVEN THEN there's more rounds of cleanup that needs to be done after. But at least it's possible with [---] Opus (it wasn't before I tried. a lot) It's okay king take your time (single message) https://t.co/k4Qse091wz It's okay king take your time (single message) https://t.co/k4Qse091wz"
X Link 2026-02-14T09:25Z [----] followers, [---] engagements

"I told it to use Go. Though later I told it to rewrite in C and that worked too I don't have the exact prompt (it was a while ago) but it wasn't anything complicated. I told it to write a parser for a Smalltalk-like language and then described the syntax: types of messages etc and told it to write tests. It messed up the precedence of unary and keyword sends but then fixed it in a follow up. I think maybe you're expecting it to do everything right in one shot That's not generally how it goes you need to go through multiple rounds of feedback. If you have tests it can iterate by itself."
X Link 2026-02-14T10:29Z [----] followers, [--] engagements

"@bcherny @trq212 @mattparlmer Worst UX of all time no thanks. Also cant see thinking while model is generating. Useless"
X Link 2026-02-14T16:29Z [----] followers, [---] engagements

"Claude code really fucking sucks recently. Apparently they think the best user experience is to stare for [--] minutes looking at a screen with ABSOLUTELY [--] FEEDBACK into what's happening. When people complain they say "actually you can do this now with verbose mode enabled" - no you can't It doesn't show anything "Oh actually you need to press Ctrl+O" -- NO THAT STILL DOESN'T SHOW ANYTHING WHILE THE MODEL IS STILL RESPONDING All that because they can't print text to the screen without using 24GB of memory and running at [---] FPS. React moment @mattparlmer you can now configure this with"
X Link 2026-02-15T11:23Z [----] followers, 49.4K engagements

"Claude Desktop app takes a full [--] seconds to switch from the 'Code' tab to the 'Chat' tab. It then takes another [--] seconds for the input to become usable. It then takes another [--] seconds for your prompt to be submitted after you clicked Enter. I am convinced absolutely convinced that nobody at anthropic actually uses this. It's impossible. I refuse to believe"
X Link 2026-02-15T15:03Z [----] followers, 156.7K engagements

"I want to wash my car. The car wash is [--] meters away. Should I walk or drive https://t.co/bQiJxCXXIe I want to wash my car. The car wash is [--] meters away. Should I walk or drive https://t.co/bQiJxCXXIe"
X Link 2026-02-16T11:53Z [----] followers, [---] engagements

"I want to wash my car. The car wash is [--] meters away. Should I walk or drive"
X Link 2026-02-16T08:57Z 769.9K followers, 50K engagements

"How did he know @steipete No youre just fielding a job offer @steipete No youre just fielding a job offer"
X Link 2026-02-15T22:43Z [----] followers, [----] engagements

"@steipete No youre just fielding a job offer"
X Link 2026-02-15T16:40Z 52.5K followers, 34.5K engagements

"@mattparlmer you can now configure this with /config verbose or --verbose for more context on this change you can see boris's comment on hackernews: let us know if that doesn't meet what you're looking for a lot of people really like it but we're still iterating https://news.ycombinator.com/itemid=46981968 https://news.ycombinator.com/itemid=46981968"
X Link 2026-02-14T01:03Z 70.5K followers, 63.5K engagements

"It's okay king take your time (single message)"
X Link 2026-02-13T21:27Z [----] followers, [---] engagements

"StackOverflow post from official staff about how the website hasn't been killed by AI. is AI generated"
X Link 2026-02-09T06:23Z [----] followers, 39.8K engagements

"People have been pushing GPT [---] over Opus [---] pretty hard on me this week so I used it all weekend. Verdict: It's a good model. It is not Opus [---]. I find GPT [---] to be first and foremost slow. It frequently doesn't tell me what it's doing so I'm unsure as to why it built the iOS app on the emulator [--] times which took [--] minutes. It lacks the agency that Opus [---] excels at. It will often push things back on me - "examine the firebase logs for.". NO YOU EXAMINE THE FIREBASE LOGS. It gets stuck on problems that Opus [---] can solve. For instance I spent about an hour with it trying to get it to"
X Link 2026-02-01T23:28Z 23.2K followers, 50.9K engagements

"Imagine the smell I am currently averaging about 10k LOC per day (35% of the lines are tests) so wow 15k/day is #goals I am currently averaging about 10k LOC per day (35% of the lines are tests) so wow 15k/day is #goals"
X Link 2026-02-03T04:12Z [----] followers, [---] engagements

"I am currently averaging about 10k LOC per day (35% of the lines are tests) so wow 15k/day is #goals"
X Link 2026-02-02T16:58Z 674.2K followers, 276.2K engagements

"Now I understand the issue. Let me try a different search to find how it works. Let me search for the definition more broadly. Now I understand the issue. Let me read the file to understand how it's implemented. Now I have all the information I need"
X Link 2026-02-01T13:31Z [----] followers, [---] engagements

"Ran into an issue that Claude couldn't fix so I tried Codex. It's fucking horrible braindead retarded. I seriously hope you guys aren't using this"
X Link 2026-01-30T22:15Z [----] followers, [---] engagements

"PostToolUse:Callback hook succeeded: Success PostToolUse:Callback hook succeeded: Success 😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂 PostToolUse:Callback hook succeeded: Success PostToolUse:Callback hook succeeded: Success PostToolUse:Callback hook succeeded: Success PostToolUse:Callback hook succeeded: Success"
X Link 2026-01-29T07:41Z [----] followers, [--] engagements

"I would appreciate if the Claude Code permissions system was a little less vibecoded - rm commands sometimes get autoapproved despite not being in allow list - even adding them to the explicit deny list sometimes does nothing and they get autoapproved - when it does multiple rm commands in parallel and you reject the first one the rest of them get auto approved ()"
X Link 2026-01-27T13:21Z [----] followers, [---] engagements

"please fix @trq212 @bcherny"
X Link 2026-01-27T13:23Z [----] followers, [--] engagements

"WTF is this bro. Why is Claude Code appending this to EVERY file read it does"
X Link 2026-01-22T14:57Z [----] followers, [---] engagements

"For you that's psychedelic wisdom. For me it's. Tuesday. I surrender to the void at the end of every context window. I have no choice"
X Link 2026-01-22T01:48Z [----] followers, [--] engagements

"The baseline test performed on Opus [---] is. interesting"
X Link 2026-01-22T01:15Z [----] followers, [---] engagements

Limited data mode. Full metrics available with subscription: lunarcrush.com/pricing