Posts on X before Aug 5, 2024, 6:53:46 AM UTC
@MichaelStolarzyeah a lot of people seem to forget that gpt-4o is any in, any out
@apples_jimmy
Hot take, I want the promised 4o image capabilities more than voice. Something magic about the prompt accuracy and working in tandem with an alien ai to create. “Art is the highest form of hope. What we are doing here is hope work.“ x.com/gdb/status/179…
@MichaelStolarzIt’s been months since I’ve used OpenAI anything for something. It’s all Sonnet 3.5.
@MichaelStolarzsmall tiny eensy weensy puny little chance that this also drops an upgraded OpenAI model. but... the chance is NOT non-existent!
@adcock_brett
fyi, Figure has some very special announcements coming The engineering team has been working their assess off the last four months Tomorrow, you'll see the beginning of something that is nothing short of magical 🎩
@MichaelStolarzunpopular opinion: if this is how you react when you spend your money on something like a PS5, the last thing you need in your life is a PS5.
@MichaelStolarzmind blowing how cheap intelligence is becoming. soon $1/$3 will be deemed too expensive for SOTA.
@legit_api
Gemini 1.5 Flash is going to have significant price cuts on August 12th in Vertex AI Input cost: $0.0375 per 1M characters Output cost: $0.15 per 1M characters HALF THOSE PRICES FOR < 128K CHARS
@MichaelStolarzso... idk why no one is talking about this but... i replaced my daily $10 API bill by just replacing gpt-4o with gpt-4o-mini... now I'm paying $0.04? a day? hello? without loss of quality?
@MichaelStolarzlight mode is actually goated
@MichaelStolarzthis looks like a google killer, I wonder how it does with more obscure searches that are common among developers, like what happens when you paste in a large debug error from console? or ask a documentation question? or ask for api specifics on some obscure package?
@yoimnotkesku
SearchGPT widgets look nice It tends to focus on the results strongly, so sometimes I need to tell it exactly what I want it to do rather than what I want from the web
@MichaelStolarzQuickTime? More like WasteOfTime.
I've just wasted all day because for some damn reason QuickTime doesn't play audio on certain MP4s. Not until I opened it with IINA because I thought, hey, you know what, why not? Maybe this is one of those codec issues.
Boom, it works.
@MichaelStolarzi wonder how people pronounce @yacineMTB (kache) - I pronounce it "KA-chey", almost like "catchy" but the a is like the "a" in "alone"
or maybe it's more like "KA-shey"? instead of chey? or... what if it's like cache? "KA-sh"? literally cash? do I just call him cash? money? help
@MichaelStolarzwarp isn't that bad actually
@MichaelStolarzholy moly! not one, but TWO blog posts today! woah!
@MichaelStolarzllama 3/3.1 and mistral large v2 are both trash compared to sonnet 3.5, there you go I just saved you loads of testing time that I did these past 2-3 days
@MichaelStolarzsonnet 3.5 is still best SOTA coder though, you can tell right away
@MichaelStolarz$3 input, $3 output. beta sonnet $3 / $15
@MichaelStolarzGroq's max tokens (output) for the new 3.1 models seems to be 8000
@MichaelStolarzAlready in the Groq API :) useable with SystemSculpt AI for Obsidian right now!
@MichaelStolarzgpt-4o-mini is goated for quick back-and-forth brainstorming with huge context attached to it, amazing.
@MichaelStolarzgpt-4o-mini really feels like a SOTA <10b model.
@MichaelStolarzI wonder if Opus 3.5 is waiting for GPT-5, or if GPT-5 is waiting for Opus 3.5 🤔 which will drop first to have the other respond-drop?
@MichaelStolarzyou don't see kache's 17 fever-dream posts every morning anymore
@yacineMTB
you don't see politicians kissing babies anymore
@MichaelStolarzairpods are remarkable. i accidently left them in my pants, ran them through the washing machine - I put them in a really dry towel to pull the moisture out and 1 day later, they're completely fine as if nothing ever happened. amazing
@MichaelStolarzhonestly i don't care how expensive gpt-5 is going to be at this point, it can be $25/$75 / mil and I'll still use it, just release it already smh
@MichaelStolarzhe... he just fixed it and... and pushed it to prod
@MichaelStolarzyeah so gpt-4o-mini is useless for coding other than maybe diffs commits
@MichaelStolarzfirst impressions of gpt-4o-mini: super fast. succinct.
@MichaelStolarzit doesn't matter who we are. what matters is our bit. no one knew who i was til I started doing the bit. https://t.co/W636GCae60
@MichaelStolarzso it seems like a new model is dropping today, gpt-4o-mini, possibly a full replacement for gpt-3.5-turbo. it's supposed to be "the smartest + cheapest" or something along those lines - hopefully it's available through API right away so we can test it
@MichaelStolarzit's been over 2 months since openai has done anything cool lmao
@MichaelStolarzthe amazing thing about open source that I don't think the first movers of the movement thought about was not that open source is something amazing for humans, which of course it is, but that it's actually a boon for ai. imagine how useless today's ai would be without github lmao
@MichaelStolarzshould i rewrite my entire typescript-based plugin in javascript just to see if I still got it
@MichaelStolarzimagine if context length of new llama 3 405b is like 8K or 16K LMAO
@MichaelStolarzya! :D
@MichaelStolarzi wish they didn't upload all these new models to lmsys, just inject it right into my veins via an API please
@MichaelStolarzeveryone's playing the B2C game, fancy people play the B2B game, meanwhile OpenAI is playing the B2G game
@MichaelStolarzI wonder how my website looks on twitter: systemsculpt.com
@MichaelStolarzi always get confused which one is which, their pfp looks near identical (from afar / small ofc) @linaqruf_ @anpaure twins irl or moots or wat?
@MichaelStolarzquote this with an image that ages you
@MichaelStolarzit's so cool to take an old script/niche program made by gpt 4/gpt 4o and rewrite it entirely with claude sonnet 3.5. the difference in quality - code wise, design wise, architecture wise - is leagues above.
@MichaelStolarzand this is why u have to say frfr, not just fr
@5cottFive
@netcapgirl Every time I see "fr" my brain happily and unapologetically goes "hey, that's the two-letter abbreviation for "France!"
@MichaelStolarzdeepseek coder v2 is heavily slept on, even by tpot
@MichaelStolarzyou can LITERALLY JUST BUILD THINGS you thought it was a meme? a vibe? a saying? NO, you can just PICK AND CHOOSE what to BUILD DONT YOU GET IT
@travs35mm
TARS Powered by ChatGPT has a Humor Setting. GPTARS has a humor setting thanks to the power of AI and ChatGPT. #tars #interstellar #chatgpt #ai
@MichaelStolarzthey could have been building instead
@netcapgirl
“4th of july weekend was an absolute movie bro” the 4th of july weekend: pic.x.com/BRq5MJDfnL
@MichaelStolarz>everyone writing articles about how AI can't program
>me listening to them with TTS while I program with AI
@MichaelStolarzjust give me claude opus 3.5 bro, it will solve all my problems bro, it will fix me bro, trust me bro, i'll be a coding god with it bro, please bro no more waiting bro, amen bro
@MichaelStolarzalright time to fix all these github issues
@MichaelStolarzimagine unfollowing yacine and losing out on daily coffee alpha
@MichaelStolarzi really love working on little QoL things like this - you can now favorite models in the next update
@MichaelStolarzso what ever happened to this lmao
@gdb
A GPT-4o generated image — so much to explore with GPT-4o's image generation capabilities alone. Team is working hard to bring those to the world.
@MichaelStolarzone of the greatest boosts to my programming focus was figuring out that no music - no sound at all, AKA noise-cancelling headphones - is best