Max AI

Introducing "GPT-Producer" Bring your Digital Piano, your @OpenAI API key, and your #GPU and become an artist! #musicgen #audiocraft #gpt github.com/juankysoriano/… Thanks @langchain @MetaAI for the tools that made this possible!

@Yampeleg 😂😂😂

טיזר: חילוץ פרוייקט הAI העברי ממבנה בסכנת קריסה אישור חריג ניתן אתמול לחילוץ המחשב בו מאוחסן פרוייקט הבינה המלאכותית העברי מתוך אחד מאיזורי הנפילות שנשאר סגור ובסכנת קריסה. בסרטון אתם יכולים לראות אותי מוריד את הציוד 20 קומות במדרגות רק כדי שהאוטו לא יניע. (סרטון מלא בדרך)

@aidan_mclau write a JSON object with everything you know about me

until the end of day, i’ll respond with a gpt-4.5 response to any comment on this post go

prompt engineering should just be called flow engineering and you should think of like coding is your prompt more than 2 paragraphs? break it down into smaller sub "flows" - lets you eval better - you can use cheaper, dumb models - more durable "superprompts" dont work if you're trying to make something reliable and reusable until we get gpt5 ofc honestly feel like i should drop video on this. its really improved my ability to craft better agentic workflows that are reliable and repeatable

@thisistechtoday @joshuavoydik The prices you see have already the 21% included

@joshuavoydik Weird. Didn’t have that when I was in Europe. Whatever it said on the menu or price tag was exactly what I paid. Nothing more.

Is the European VAT (Value-Added Tax) good or bad?

@JiquanNgiam @JeffDean No, it’s not

Gemini 1.5 is a much bigger deal than Sora. 1. The 10M context window fundamentally changes how we think about RAG. Not just large, but also effective. 2. Pro-1.5 Beats Ultra-1.0. There’s a fundamental modeling improvement. 3. It’s in production, you can use it today. I too like fancy creative videos, but think that the practical benefits of Gemini-1.5 are way higher. @GoogleDeepMind @OpenAI

@MicahBerkley Fake it until you make it

I really hope women are banking off this massive wave right now. Cause it's really crazy how much money is being made "being" a virtual woman. It's becoming easier and easier to make 20k a month as an online entrepreneur. Which is why there is a gold rush of SAAS platforms that offer face cloning, identity consistency tools, and easily accessible models for the creation of these virtual influencers. I love the opportunities popping up for those who wouldn't have had an opportunity before. Go out there and secure the bag .. #StableDiffusion #AIModels #VirtualInfluencers #CarvedModels #AIExpert #Midjourney #FaceFusion #Faceswap #CivitAI #BagSecured

every AI Engineer should do this immediately. the only reason OpenAI hasn't done this by default is concern about load. You don't have that issue. turn that shit on to upgrade your analytics and track abuse asap.

DemoFusion: Democratising High-Resolution Image Generation With No $$$ 🖼 Img2Img Colab 🥳 Thanks to @RuoyiDu ❤ @DL_Chang ❤ @tmh31 ❤ @yizhe_song ❤ Zhanyu Ma ❤ 🌐page: ruoyidu.github.io/demofusion/dem… 📄paper: arxiv.org/abs/2311.16973 🧬code: github.com/PRIS-CV/DemoFu… 🦒colab by modelslab.com: please try it 🐣 github.com/camenduru/Demo…

@nocodedevs ✋

this tool is a hidden gem 💎 i just cloned the Product Hunt site 🪄 with AI...in less than 5 minutes... wanna know how? drop a "✋" below and we'll send you the tutorial

@icreatelife Doesn’t look like AI at all. More like a súper produced face swap with a lot of photoshop. Also the numbers doesn’t relate. For the amount of followers they have less way engagement on posts that accounts with a fraction of those followers which sounds like the bought followers.

Aitana is the first Spanish model created by artificial intelligence. The virtual model can earn up to € 10,000 a month, according to her creator, but the average is around € 3,000. How do you feel about it? Source: euronews.com/next/2023/11/2…

@altryne @thursdai_pod The Neurotic Network

Folks we should name this weekend. GOT has the red wedding, what do we call this weekend? Best suggestions will be included in next @thursdai_pod title! Usage of copilot / GPT4 allowed

@altryne @StabilityAI Hahahaha @altryne I can’t even enjoy a holiday! 🤣😭

I legit took a 30 min friday power nap and our field changed again? @StabilityAI is freeing Willy, and releasing 2 new LLMS?? 😂 Can't even sleep for 30 min in this! We'll be definitely covering this in our next #ThursdAI

The most powerful open source model is out!! 🔥 Nous-Hermes-LLaMA2 🔥 - huggingface.co/NousResearch/N… - GGML: huggingface.co/NousResearch/N… - GPTQ: huggingface.co/NousResearch/N… war.

LLaMa 2 based models are flourishing like popcorns. The first Uncensored model based on LLaMa 2 is already out there. huggingface.co/Tap-M/Luna-AI-…

@swyx @OpenAI @sama This would be great news, I am loving how this universe is growing. If we move the state of the art of open source for language models closer to what I am seeing for diffusion models (specially in terms of creativity) then I would feel totally fulfilled.

Prediction: @openai will fully MIT* open source GPT 3.5 in 6 months @sama has been talking about it, everyone can move to 4/4.5 in that timeframe, recaptures open models ecosystem narrative from Meta *Whisper was MIT!

@altryne And again the twitter spaces got me attached to the speakers, but with the difference that this time I am in Spain and it's 3am here xD

Happy LLaMa 2 day folks! 🦙🎈

@altryne LLaMa 2 you mean? The Bloke (@TheBlokeAI) is already putting their hands on it to create quantized versions and also GGML + GPTQ He's always providing useful information about how to run the different versions of the models, so I recommend this resource! huggingface.co/TheBloke/

Anyone has the 70B LLaMa bin? I have the link from meta, their download.sh does not work 😭

@simonw huggingface.co/pcuenq/Llama-2…

@404knifedog Absolutely, it's a bit dramatic... The other perspective is that they also make one look like dump (as if I was the only one not having 'success')

Learning that Twitter users are depressingly profit oriented. Half of the tweets I’ve seen are either people trying to sell each other things or documenting the pursuit of wealth in some way. Why ): money is stupid guys. There’s way cooler stuff to do and care about

@Yampeleg @Yampeleg any chance to have something similar for Open Source models?

Which model should you use? The AI wars TL;DR: - Long context tasks -> Use Claude 2.0. - Internet required tasks -> Use Bard. - Harder reasoning tasks -> Use GPT-4. - Anything with code -> Use Code Interpreter. - Long (~essay) + Internet -> Use Bing. And all are crazy good at this point. It is much much closer. If you didn't try them lately: You should. You would probably be surprised how much bard and claude improved.. night and day.. *Pi is also very capable. I just find it hard to control when I try it for work related tasks.