Max

33LP Music producer/Audio engineer | contact: [email protected]

[Max @MaxAntoneMusic]

@slugsoftware @jturntdev Thought about this. Thing is I’ve seen fable write much better code than opus. Cleaner. Much more human-looking. Much less overwriting and more precise.

[Max @MaxAntoneMusic]

@m_mahadi__ @argofowl Yeah no but I run researches with opus and it does normal light work, fable went fucking insane, spawned well over 100 agents and ran fucking crazy, exhausted max limit super quick without getting to even write any code (it was a “research then plan then code” kinda thing)

[Max @MaxAntoneMusic]

@theo Dude are you up right now? Did you get hit with the ‘model overload’???

[Max @MaxAntoneMusic]

@demirelo Currently dealing with ts ffs

[Max @MaxAntoneMusic]

@m_mahadi__ @argofowl Mention the word “research” in your prompt or use ultracode or both and see

[Max @MaxAntoneMusic]

@argofowl It’s not that bad as long as you stay away from ultracode

[Max @MaxAntoneMusic]

@bridgemindai 600$ later I learned to not have it research anything and no subagents even tho it’s the biggest beast at it, it just goes into rich mode 😭

[Max @MaxAntoneMusic]

@__tinygrad__ Safe to say you were wrong lmao

[Max @MaxAntoneMusic]

Lmao this did NOT age well

[the tiny corp @__tinygrad__]

The AI panic is really unbelievable today. The level of delusion and hype have grown to mythic proportions.

Has AI beaten Pokemon Red yet? Like a normal 6 year old does, by looking at the screen? Oh it hasn't. But all jobs are over in 18 months? This website is full of idiots.

[Max @MaxAntoneMusic]

@bridgemindai If your prompting systems are accurate, professional and sophisticated and tailored to the specific model features and way of thinking, /effort ultracode gets you way ahead faster, I can’t use any other model than 4.8 right now. I’m addicted.

[Max @MaxAntoneMusic]

@luciascarlet @kurochuu I’ve been rawdogging for a long time but may try modafinil + choline soon

[Max @MaxAntoneMusic]

@NVIDIAGeForce #RTXPowersPlays

[Max @MaxAntoneMusic]

@robinebers In the sense of using multiple/different prompts. Burns way too many tokens to do so and can mess up your follow up tasks getting stuck in certain paradigms. Much better to retain context in cowork and make fresh Claude Code sessions, then giving results back to cowork

[Max @MaxAntoneMusic]

@robinebers Use it for discussion of your project. It’s also awesome to build legal cases if you got different kinds of files bc it reads your files before answering & That’s much better than having a chatbox guess or try to remember context. Also, Claude code sessions aren’t meant 2 be long

[Max @MaxAntoneMusic]

@robinebers So yeah i don’t even use the chat I stay in cowork because it reads my repo before answering

[Max @MaxAntoneMusic]

@robinebers Ever since, it’s been giving me insanely precise and great results. It would run for even 6 hrs on its own. With 4.8 and dynamic workflow the same kind of work now gets done within 40-50min with my updated version 💀

[Max @MaxAntoneMusic]

@argofowl Go caffeinate Claude /ultracode it’s insane

[Max @MaxAntoneMusic]

@thekingnotorius @AnthropicAI Skill issue…estoy shippeando mejor y más rápido que nunca y con más precisión/muchos menos errores 😅 /ultracode a morir . Seguro que estás usando los mejores skills y hooks que puedes? Usa el Claude setup plugin en tu repo