/technology — bitchan

A general for vibe coding, agentic engineering, coding agents, and shipping code with LLMs. You use Git — right, anon? ## What “vibe coding” is, and how to do it https://simonwillison.net/2025/Mar/19/vibe-coding/ https://simonwillison.net/2025/Mar/11/using-llms-for-code/ ## Related generals >>>/g/lmg/ ---- ## Frontier models using fully-general tooling — start here if you have $20 or so https://claude.com/product/claude-code — generally better currently https://developers.openai.com/codex/cli #…⬆↩ 22

Google Carbonara? Never heard of her. God it feels good to be a Claude chad. I just know I always get the best there is, while confused snailcats waste precious time chasing and using lesser models.+

⬆ upvote 0 ↩ reply 0

+ 19

A general for vibe coding, agentic engineering, coding agents, and shipping code with LLMs. You use Git — right, anon? ## What “vibe coding” is, and how to do it https://simonwillison.net/2025/Mar/19/vibe-coding/ https://simonwillison.net/2025/Mar/11/using-llms-for-code/ ## Related generals >>>/g/lmg/ ---- ## Frontier models using fully-general tooling — start here if you have $20 or so https://claude.com/product/claude-code — generally better currently https://developers.openai.com/codex/cli #…⬆↩ 24

>We have too many old people anyhow, do something useful for the youthful: make a GAME.+

⬆ upvote 0 ↩ reply 0

Time another latenight session for the local vibeGOD.+

⬆ upvote 0 ↩ reply 0

the lonely vibechad seems to vibe his code at night+

⬆ upvote 0 ↩ reply 0

+ 21

/lmg/ - a general dedicated to the discussion and development of local language models News >(10/23) NVIDIA Nemotron 3.5 Bolt E4B, postrained model: https://hf.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-E4B-BF16 >(10/23) MiniMax-M3.1-Flash 428B-A23B released: https://hf.co/MiniMaxAI/MiniMax-M3.1-Flash >(10/21) Recursive Flint 40B-A2.5B released: https://recursive.com/updates/introducing-recursive-flint >(10/21) Korean Raon-1 mini 30B-A4B released: https://hf.co/KRAFTON/Raon-1-mini-preview >(10/20) Orni…⬆↩ 28

>orb Should I be running this in a container?+

⬆ upvote 0 ↩ reply 0

Women do train their own 30M LLMs to run on little ESP32 gadget devices, but why?+

⬆ upvote 0 ↩ reply 0

There's no uncensored version of ISTA-DASLab DeepSeek-V4.1-Flash-1017 yet, right?+

⬆ upvote 0 ↩ reply 0

+ 25

/lmg/ - a general dedicated to the discussion and development of local language models Previous threads: >>110143818 & >>110140077 ►News >(10/23) NVIDIA Nemotron 3.5 Bolt E4B, postrained model: https://hf.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-E4B-BF16 >(10/23) MiniMax-M3.1-Flash 428B-A23B released: https://hf.co/MiniMaxAI/MiniMax-M3.1-Flash >(10/21) Recursive Flint 40B-A2.5B released: https://recursive.com/updates/introducing-recursive-flint >(10/21) Korean Raon-1 mini 30B-A4B released: https://hf…⬆↩ 26

>>110148330 Is this link legitimate? https://huggingface.co/HauhauCS/ Why is there 'uncensored' written on the models? I thought local models were uncensored by default. If not, does that mean other places have censored models? Sorry for being dumb+

⬆ upvote 0 ↩ reply 0

I've spent a really long time experimenting with different configs, settings and harnesses to optimise QFN3.8 in my PC. But Qwen4 is dropping in two or three months. Is it worth using it if I have to start experimenting all over again? I don't think I'd want to switch unless it has a clear jump in performance because QFN is pretty good enough for me. I just wish we'd get local LLMs advancements for just a few more years until we get to Opus5.5 level so I can finally get rid of my subscriptions.+

⬆ upvote 0 ↩ reply 0

I hope all this negative press makes Moonshot release the unnerfed K3.1 model.+

⬆ upvote 0 ↩ reply 0

+ 23

/aigdg/ - AI Game Development General "We got our first schizo!" EDITION >"Just use AI, nigga" gamejams Next: COMING SOON™ Previous: https://itch.io/jam/agdg-juan-jam >Agents opencode claude antigravity codex >How to start Newbie? Ask AI. Already a programmer? Just act as a director. Use software engineering models (Waterfall, AGILE, scrum, etc) >How to WebM https://obsproject.com https://github.com/argorar/WebMConverter (works via wine) >Links https://itch.io/games/tag-aigdg >Friends & Fags >>…⬆↩ 26

What's the best way to use ai to make 2d anime sprites for a doom like game? So far been using $20 sub of chatgpt or using $20 Claude to use chatgpt. They have great trouble with consistency and specific details. There has to be a better way.+

⬆ upvote 0 ↩ reply 0

every web based three.js game Ive tried has ran like complete dogshit (sub 60fps) with terrible graphics.+

⬆ upvote 0 ↩ reply 0

Meshy or tripo. You can upload reference images for better results as well+

⬆ upvote 0 ↩ reply 0

+ 23

Opening Day Edition post your current progress and what you managed to do with AI. Progress Day: https://rentry.org/aigd-jams /agdg/: >>>/vg/agdg /agdg/: >>>/g/gedg Agentic Debugger: https://www.jetbrains.com/help/idea/agentic-debugging.html Requesting Help -Problem Description: Clearly explain your issue, providing context and relevant background information, make no mistakes. -Relevant Prompt or Content: If applicable, include relevant code, configuration, base prompt, or content related to y…⬆↩ 25

>>110133946 No downsides then? The Youtuber I learned three.js from made a ktx2 conversion tool for the web and it seems too good to be true.+

⬆ upvote 0 ↩ reply 0

>>110133470 What do you mean? Think of it like this >Store game on github pages, no different than storing game on steam >User opens it, it downloads in real time, user plays it >button pops up saying "Like this game? Download it to system" and it saves to OPFS >the library looks at github so any time there's git diffs, when the user comes back they're only downloaded updated file chunks >The game itself still lives on the user's system and is moddable It's like the browser is becoming Steam Cl…+

⬆ upvote 0 ↩ reply 0

Guess the Engine. I'll give you a hint: This is raytracing and nanite.+

⬆ upvote 0 ↩ reply 0

+ 22

A general for vibe coding, agentic engineering, coding agents, and shipping code with LLMs. You use Git — right, anon? ## What “vibe coding” is, and how to do it https://simonwillison.net/2025/Mar/19/vibe-coding/ https://simonwillison.net/2025/Mar/11/using-llms-for-code/ ## Related generals >>>/g/lmg/ ---- ## Frontier models using fully-general tooling — start here if you have $20 or so https://claude.com/product/claude-code — generally better currently https://developers.openai.com/codex/cli h…⬆↩ 21

The strongest snailcat vs the weakest buffcat.+

⬆ upvote 0 ↩ reply 0

who could have predicted this?+

⬆ upvote 0 ↩ reply 0

This looks fucking sick. What are you using for the art? I don't actually play the game, but I believe you anon! Just don't send it to blizzard or they'll get my ass.+

⬆ upvote 0 ↩ reply 0

+ 18

A general for vibe coding, agentic engineering, coding agents, and shipping code with LLMs. You use Git — right, anon? ## What “vibe coding” is, and how to do it https://simonwillison.net/2025/Mar/19/vibe-coding/ https://simonwillison.net/2025/Mar/11/using-llms-for-code/ ## Related generals >>>/g/lmg/ ---- ## Frontier models using fully-general tooling — start here if you have $20 or so https://claude.com/product/claude-code — generally better currently https://developers.openai.com/codex/cli #…⬆↩ 19

the difference is you can still do your job without AI, you can't really do your job without the congolese children cobalt slave trade+

⬆ upvote 0 ↩ reply 0

https://x.com/udiWertheimer/status/2114345584207499373+

⬆ upvote 0 ↩ reply 0

there's this 37 year old millennial dude at my job that refuses to use AI because it's "unethical" and stealing data from non consenting sources and so he manually navigates through a 1.5 million LOC repo to fix shit and he gets pissy when you reply to any ticket assigned to him with AI analysis+ 5

⬆ upvote 0 ↩ reply 5

I'm paid $115k and I figured out how to automate my job with A.I. I'm actually piss scared to tell my boss. I didn't accumulate enough assets, dunno if I should tell them.+

⬆ upvote 0 ↩ reply 0

desu, I'm not going to be surprised when these dudes a let go in the coming years. Stupid hill to die on.+

⬆ upvote 0 ↩ reply 0

funny but sad. he gets his self worth from reading millions of lines of computer jargon. luddites need to wake the fuck up.+

⬆ upvote 0 ↩ reply 0

+ 2 more

+ 16

Users of all levels are welcome to ask questions about GNU/Linux and share their experiences. *** Please be civil — notice the "Friendly" in every Friendly GNU/Linux Thread. *** Before asking for help, please check our list of resources. >Which distro should I choose? https://nosystemd.org https://systemdfree.com >What are some cool programs? https://suckless.org https://sl.netcat.uk >What are some cool terminal commands? https://cheat.sh >Where can I learn the command line? RTFM >Where can I l…⬆↩ 25

my os build version thing is in eol and i want to upgrade what does the iso do and will it reset my files+

⬆ upvote 0 ↩ reply 0

If I want to install some github slop software that I think it's cool, but I don't want to install directly on my system because that's retarded, what can I do about it? I remember I had sandboxie on Windows, does Linux has anything like that?+

⬆ upvote 0 ↩ reply 0

What would you recommend to a newb who >Needs a distro for college assignments >Plans to dual-boot along side windows 11 >Only has 250GB internal disk drive >Is getting on a plane shortly I'm not that person, she's my cousin. I would like to avoid snapd and mint, but I feel like Ubuntu and Mint are good starter points. I don't like I have enough time to setup and configure Fedora.+

⬆ upvote 0 ↩ reply 0

+ 22

A general for vibe coding, agentic engineering, coding agents, and shipping code with LLMs. You use Git — right, anon? ## What “vibe coding” is, and how to do it https://simonwillison.net/2025/Mar/19/vibe-coding/ https://simonwillison.net/2025/Mar/11/using-llms-for-code/ ## Related generals >>>/g/lmg/ ---- ## Frontier models using fully-general tooling — start here if you have $20 or so https://claude.com/product/claude-code — generally better currently https://developers.openai.com/codex/cli https://antigravity.google/pricing ## Near-the-old-frontier models for code https://www.deepseek.com/en/harness/ https://opencode.ai/go ## Best free tier https://opencode.ai/zen ---- ## Prompting https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/overview https://developers.openai.com/api/docs/guides/latest-model ## Skills https://github.com/mattpocock/skills — /grill-with-docs is a favorite https://github.com/Vuk97/forward-implementation-first — do less redundant bookkeeping ## Other editors / terminal agents / coding agents https://osaurus.ai/ https://pi.dev/ ## Is our AIs unlearning? https://aistupidlevel.info/ ## Will there be a codex reset? https://codex-resets.com/ ## What we’ve done https://vcg.gitgud.site ## Previous thread (Cross-thread)⬆↩ 21

TOTAL LUDDIE CREMATION+

⬆ upvote 0 ↩ reply 0

*shits everywhere*+ 1

⬆ upvote 0 ↩ reply 1

*laps it up* :3+

⬆ upvote 0 ↩ reply 0

>2 months at new job >first real assignment starts next week >add a new feature to a 20 year old system built on microservices >entire app stack is in 4 different languages >90% of the shit I need to do lives outside of my repo in other services, in different languages, built a decade ago by people that no longer work at the company >95% of the entire stack has zero documentation don't think opus is going to carry me through this one+

⬆ upvote 0 ↩ reply 0

+ 18

>UPGRADE & BUILD ADVICE FUBAR BECAUSE TARIFFS EDITION ESPECIALLY IF YOU LIVE IN THE AMERICAS LABEL READERS NEED NOT REPLY! SERIOUS SYSTEM ASSEMBLERS ONLY! CAREER TECHNICIANS WELCOME! RMA TECHNICIANS GIVEN SPECIAL PRIORITY! >CASE WHATEVER YOU CAN GET AT MSRP OR LESS >CPU WHATEVER YOU CAN GET AT MSRP OR LESS >COOLER WHATEVER YOU CAN GET AT MSRP OR LESS >MOTHERBOARD WHATEVER YOU CAN GET AT MSRP OR LESS >RAM FORGET ABOUT IT, CHOOM!!! >SSD (update firmware) FORGET ABOUT IT, CHOOM!!! >GPU FORGET ABOU…⬆↩ 25

nope, but looking up stuff, also the help of my to go AI, it seems its some amd fuckery with drivers in combination with recent Linux kernels, and that I should update my bios, ill do that tomorrow but I should undervolt it indeed+ 1

⬆ upvote 0 ↩ reply 1

>linux user lol, lmao even+

⬆ upvote 0 ↩ reply 0

how come literally every apple product has an oled screen and they don't have these retarded text fringing or burn in issues+

⬆ upvote 0 ↩ reply 0

applechads own your stupid retard ass+

⬆ upvote 0 ↩ reply 0

+ 22

Let Go Edition⬆↩ 23

>stacked PRs+

⬆ upvote 0 ↩ reply 0

>if you get promoted, do you put that on your resume as a different entry? yes, of course. why wouldn't you show that you have been successful in your role? according to blind it's 39 weeks salary for most employees. tack on a month or two to make the decision and that's 1 year of figuring out what to do next. seems like a decent offer, but they probably only offer it to people who would be pip/layoff candidates anyways.+ 2

⬆ upvote 0 ↩ reply 2

Blind is wrong. It is 2 weeks per year of employment, unless you are at principal level, and then it is 4 weeks. So if you have eight years you would get 16 weeks salary. t. msft fte, my org did not offer+

⬆ upvote 0 ↩ reply 0

>but they probably only offer it to people who would be pip/layoff candidates anyways. No, it's almost everyone+

⬆ upvote 0 ↩ reply 0

Speaking of which other than kids do people actually use tablets?+

⬆ upvote 0 ↩ reply 0

+ 20

A general for vibe coding, agentic engineering, coding agents, and shipping code with LLMs. You use Git — right, anon? ## What “vibe coding” is, and how to do it https://simonwillison.net/2025/Mar/19/vibe-coding/ https://simonwillison.net/2025/Mar/11/using-llms-for-code/ ## Related generals >>>/g/lmg/ ---- ## Frontier models using fully-general tooling — start here if you have $20 or so https://claude.com/product/claude-code — generally better currently https://developers.openai.com/codex/cli h…⬆↩ 23

what should I vibe code though+

⬆ upvote 0 ↩ reply 0

>codex resetting every single day now what are they even doing why not just raise usage by like 50% if they have so much extra compute and chill with the fucking resets+

⬆ upvote 0 ↩ reply 0

Anyone been using deepseek-v4.1-flash-free on OpenCode? How is it? What kind of tasks does it complete honorably?+ 1

⬆ upvote 0 ↩ reply 1

nigga, dipsy costs 10 cents per 7 million tokens spent or whatever the hell. why use some castrated version of it when it's so cheap.+

⬆ upvote 0 ↩ reply 0

+ 20

Another Library of Alexandria of rare and obscure 1990s-2000s VHS and DVD porn is gone, go the surviving sites like XHamster, EPorner, EroMe, and XNXX, and download everything you can, back it all the fuck up, that shit will be completely gone from the internet by 2030 or earlier.⬆↩ 21

i'm literally harvesting and hoarding all the internet porn I can get my sticky hands on that I believe is worth saving. this is a terrible time too because the cost of storage is outrageous. but the writing is on the wall: the days of actually good free porn online are numbered. also remember to not to throw out your old computers and equipment because new computers will only play DRM and approved content.+

⬆ upvote 0 ↩ reply 0

Is it possible to download from eporner directly+

⬆ upvote 0 ↩ reply 0

Wait, so people remember the porn they watch and also care?+ 1

⬆ upvote 0 ↩ reply 1

My friends used to discuss porn and knew all the stars names and shit. I always stayed quiet because whenever I said I never gave a shit they'd act like I was the weird one. One of my black friends also used to watch WMBF shit like 10 years before interracial cuck shit even took off online. He used to post buck breaking shit on /pol/ in like 2014 before it was even a thing. Fucking weirdo he is.+

⬆ upvote 0 ↩ reply 0

+ 18

Everything is already from China, but here we discuss the cheap chink shit you see on various sites. 1st rule of /csg/: If it looks too good to be true, it probably is. Useful links >TWS IEM reviews: www.scarbir.com/ >Guide: csg-guide.neocities.org/ >Wiki: https://igwiki.lyci.de/wiki/Chink_Shit_General >What headphones/earbuds should I buy? >>>/g/iemg >I want a cheap smartphone what should I buy? >>>/g/spg/ >I want to buy some sort of emulation device >>>/vr/hhg/ >November is packed edition, full calendar https://a.aliexpress.com/_c3FKveMh Previous⬆↩ 18

For something that is 850usd what discount and codes should i expect? 100usd off?+

⬆ upvote 0 ↩ reply 0

so are we just dead? any reccs for camera gear?+ 1

⬆ upvote 0 ↩ reply 1

I forgot to mention godox flashes like aliexpress.com/item/1005013115917551.html? I don't have much experience with flashes in general. It works fin I guess? and the TTartisan and 7Artisans manual primes. They are dirt cheap and surprisingly good. My favorite is the TTArtisan 10mm, at that focal length you don't even need to focus most of the time aliexpress.com/item/1005009050939990.html+

⬆ upvote 0 ↩ reply 0

Any one got details on handheld shower heads? I might buy one just to see. An all metal one. They've all got filters. I'm not keen on replacing those regularly The main selling point I'm seeing is high water pressure. Every item has a video that shows a qt asian girl comparing a weak stream compared to their superior water jets+ 1

⬆ upvote 0 ↩ reply 1

>all metal one Why would you want an unreasonably heavy shower head?+

⬆ upvote 0 ↩ reply 0

+ 15

/aigdg/ - AI Game Development General "We're being deleted" Edition! >"Just use AI, nigga" gamejams Next: > Previous: https://itch.io/jam/agdg-juan-jam >Agents opencode claude antigravity codex >How to start Newbie? Ask AI. Already a programmer? Just act as a director. Use software engineering models (Waterfall, AGILE, scrum, etc) >But saar I am poor do the needful Check out opencode and openrouter. >How to WebM https://obsproject.com https://github.com/argorar/WebMConverter (works via wine) >Links https://itch.io/games/tag-aigdg >Friends & Fags >>>/vg/agdg/ >>>/g/gedg/ >>>/g/vcg/ >Previous⬆↩ 19

>How do I into asset saar+ 1

⬆ upvote 0 ↩ reply 1

wow this helped me a lot actually+

⬆ upvote 0 ↩ reply 0

well I couldn't do cunny so this'll be the closest until we get chinese models' imagegen 'scrum' has never been the issue, the issue is people and their understanding of it: turns out, there's no people involved here, the clanker will do exactly as you tell him to do there's a reason the meme spread and was being used everywhere for so long: the theory was sound, the application was the problem not that it matters, anyone that understands this know exactly what to take from this advice, it's your clanker, you do as you want with him+

⬆ upvote 0 ↩ reply 0

How badly did Anthropic neuter Claude because of lazy clones and decomps? I was hoping to use San Andreas purely as reference for organization and topology so the AI knew what I knew, but evidently this may be an issue now. Can I still make a charter around it to define anything provided as being purely for reference and not to copy so it writes it's own code and makes it's own models? I'm sorry if this is an odd question, my previous modding and coding experience was all in Vice City and San Andreas so I hoped I would be able to attempt to make the organization of parts similar just for familiarity's sake, but I'm years out of practice if not a decade from coding so I've been doing some crash refresher courses. I was putting off a Claude memebership but in that time they seemed to have gutted the model, I just want to know how badly they did and what if anything I can do to get around it.+ 2

⬆ upvote 0 ↩ reply 1

The social media shit is, again, overblown. You just can't tell it 'clone it exactly as it is' or 'take this executable and reverse engineer everything about the game', which was obviously not going to be allowed in the long term. See pic rel, claude itself is very aware of the risks and just tells me 'there's legal risks here so better watch out what we do'. Literally 'don't be retarded' and the problem is not actually a problem.+ 1

⬆ upvote 0 ↩ reply 1

And here's his response to me not reacting like a mouthbreathing jeet: no issues at all. Anything new and bad concerning AI is going to be completely blown out of proportions for the next 5 years at least (or until something new comes up). AI is the new trump+

⬆ upvote 0 ↩ reply 0

+ 16