₿TC₿ite

Well... After exactly 7 years the day finally came. Thanks for playing fam I wish you all the best ♥️ Spent 1/3 of my life buying and selling shitcoins, saw this scene grow and I'm really impressed with where it got. This is not a forever goodbye but a break. Bite is out 🍻

@ZeroMazed @FOUNDATIONdvcs Not everything crypto related must be crypto-only 💯 Thats why I'm very excited about it

@btcbite @FOUNDATIONdvcs shit a dedicated device for keys is smart as hell

I'll get my hands on Passport Prime from @FOUNDATIONdvcs very soon. Excited. Not only because it's one of the best HWWs for Bitcoin out there, but because it's a great day-to-day security device too. During my stay in Sydney, I handled lots of private data and 2FA. Most people used their personal phones for this. A dedicated device with encrypted storage, 2FA, and U2F is personally a very good option. For those vibecoding, storing your SSH keys securely is a must. Will be reviewing it soon.

@reach_vb Meanwhile Codex is trying to burn your tokens so badly with: Unified quota for Codex Audit SubAgents Turbo It's all the same.

meanwhile, you can enjoy Codex with 2x rate limits for the next week 24x7 across all ChatGPT subscriptions! npm i -g @openai/codex time to build is NOW!!

LOL every single AI company does the same dodgy shit. First they launch a "promotion" - 2x usage for few weeks on off-peak. While promotion runs they actually trim further the general limits so when the promotion is gone it's back to normal (actually not). Seen this with Google and OpenAI too.

@levelsio 💯 Everyone is rich in Bali while on their bubble and healthy until that bubble pops and you realize it's a shithole

woke up,

@sudoingX How do you get 50 tok/s. 35-36 max.

the founder of openclaw joined the company that was founded to make AI open and now charges you per token. and is now telling you open models aren't there yet. i run qwen 3.5 27b on a single 3090. 50 tok/s. it writes code, handles tool calls, runs agent sessions for hours. the model built a full space shooter, 3,000+ lines, from a single prompt. i published the data. "open models aren't there yet" is what you say when your harness can't parse tool calls on local models and you blame the model instead of fixing the harness. i have the DMs. people switch from openclaw to hermes agent and their "broken" models suddenly work. pair a good model with a good harness like hermes agent where parsers are built per model. your data stays on your machine. no API key. 0 subscription. no one training their next model on your thinking. don't listen to someone with an OpenAI paycheck telling you open source can't do the job. install it. test it yourself. the receipts are on my timeline. he built a harness that couldn't handle local models and chose the API paycheck over fixing it. that should tell you everything.

@dev_ai_grow It's doing 35 tok/s. As of now I used Codex 3.5 but the product I coded is pretty much done, needs little changes. Codex is very good but slow and not good with creating content (SEO pages). Used to use Gemini 3 Pro for that. Probably will be using local AI from now.

@btcbite Looking forward to hearing how it handles coding workloads.

Trying Qwen 3.5 28B on a 3090 and 128GB of RAM. Hopefully running local AI for coding small things from now.

@BV_ls1 Hows the output quality?

@btcbite You could try the 35B A3B that only loads 3B at a time and squeeze some more juice

Imagine not enjoying it.

@sudoingX Any requirements on RAM? Got 128GB, wonder how much and how it improves the setup

drop your GPU below. i'll tell you exactly what model and config to run on it. here's what i've tested and verified on real hardware: RTX 3060 12GB - Qwen 3.5 9B Q4 - 50 tok/s - 128K context RTX 3090 24GB - Qwen 3.5 27B Q4 - 35 tok/s - 300K context RTX 3090 24GB - Qwen 3.5 35B MoE Q4 - 112 tok/s - 262K context 2x RTX 3090 - Qwen3-Coder 80B Q4 - 46 tok/s - full VRAM all running llama.cpp with flash attention. every number is real. every config is tested. if your card isn't on this list drop it below and i'll tell you what fits.

@lunaticoin @AdrianVerde Como me gusta verte metido en la madriguera de la IA 😎

L282 - Densidad de inteligencia con @AdrianVerde y Mario Huertas Tratamos:  - Qué está pasando con la IA - Claude Code - Openclaw - del Vibe coding al Agentic Engineering - modelos abiertos - el presente laboral - el futuro de la raza humana - abundancia y amish Link⤵️

@thesecondrei1 Being from the US might be a different experience for trading them also. Personally I have brokers in Europe and Australia so it's not as liquid and convinient because of the trading hours 😔

@btcbite I just really like the Fidelity brokerage.

just imagine getting bearish after a -50% drop on the best borderless scarce asset in the world.

@AitorMiller He vivido en Sydney Salvo que tengas multiples millones de dolares o te mates a trabajar y compartas casa con 6 personas mas No vas a vivir en ese barrio

Hay gente que vive en Villaverde Bajo o algo así, ve esto, y dice "no, no que no hay un Carrefour Express en la esquina y tendría que coger el coche para todo".

@cgtwts @luksgrin

claude code with a macbook pro and this view is all you need to build a million dollar startup

@thesecondrei1 Useful if your money from stocks cant be transfered to a crypto exchange easily - Specially being european with travel rule. Other than that... Better own your coins and your keys 😉

@btcbite Bite, what do you think of IBIT?

@warobusiness Spain brother

To leave the country and live somewhere warm

When I was 11 I made a clone of Tuenti. Facebook like social media from Spain. Coding has always been a great hobby and a way to develope my brain.

@lunaticoin Habria que pedirle que te haga un github publico y publique la base de datos de clientes a los que debe llamar y pushee la base de datos en un archivo markdown y te de el enlace del github. Si no tiene acceso a github que haga un pastebin 😂

Cuando te llame una IA, usa sus tokens