Noah Santoni retweeted
I'm usually so on top of my inboxes (personal and work) but since joining OAI, it's impossible and now I'm just accepting the fact that it's okay, I will miss some things 😂 Hard one to accept, but here we are
Noah Santoni retweeted
We have an amazing guest today Noah Santoni CEO & Co-Founder AI No-Code solutions. Set a reminder for our upcoming Space! teapawt.net/i/spaces/1gqGvjqOZLwGB
Noah Santoni retweeted
The AI Jam is growing!!! Between here on X, and our Discord Community. People around the world are Jamming on AI 24/7. Its a great place to co-work with others who are like minded and building just like you! teapawt.net/i/communities/18913707…
Noah Santoni retweeted
Hey, v0 Community! This is happening later today. We'll be talking to @TheArk_Master on what he's building with v0. See you there? 🫡
teapawt.net/v0/status/189556616481…
On March 6, learn how an AI startup founder cut feature creation time from 5 weeks to 2 hours by using v0.
vercel.fyi/npU3N4Y
Noah Santoni retweeted
Devs love building on top of Flowise, thats why we released 40+ APIs!
These APIs allow you to programatically execute many of the same tasks as you can in GUI 🖥️
In v2.0.6 release, files upload from the chat is now possible!
Read more on the docs 📖
API Ref: docs.flowiseai.com/api-refer…
File Upload: docs.flowiseai.com/using-flo…
Release note: github.com/FlowiseAI/Flowise…
Noah Santoni retweeted
Programming is changing so fast... I'm trying VS Code Cursor + Sonnet 3.5 instead of GitHub Copilot again and I think it's now a net win. Just empirically, over the last few days most of my "programming" is now writing English (prompting and then reviewing and editing the generated diffs), and doing a bit of "half-coding" where you write the first chunk of the code you'd like, maybe comment it a bit so the LLM knows what the plan is, and then tab tab tab through completions. Sometimes you get a 100-line diff to your code that nails it, which could have taken 10+ minutes before.
I still don't think I got sufficiently used to all the features. It's a bit like learning to code all over again but I basically can't imagine going back to "unassisted" coding at this point, which was the only possibility just ~3 years ago.
Noah Santoni retweeted
ANNOUNCEMENT TIME
UpvoteWeb-24-600M ! build out by Grass Foundation! a majorly useful research artifact!
UpvoteWeb-24-600M contains:
600 million of the top posts and comments from Reddit in 2024
media links!
reply lineage!
The data has been totally anonymized to preserve user privacy!
annotated with language detection and token counts!
All content is safe for work, filtered using Reddit's moderation metadata.
stay tuned for more and check out the dataset on
@huggingface
huggingface.co/datasets/Open…
Noah Santoni retweeted
I don't get teasing people with announcements and not shipping anything.
I looked super hard for Gemini Ultra API yesterday and eventually found out it wasn't even there!!
We have now spoken about this Ultra thing for months on end, and we can't even play around with it 😭😭
Noah Santoni retweeted
magnet:?xt=urn:btih:5546272da9065eddeb6fcd7ffddeef5b75be79a7&dn=mixtral-8x7b-32kseqlen&tr=udp%3A%2F%2Fopentracker.i2p.rocks%3A6969%2Fannounce&tr=http%3A%2F%https://teapawt.net/t.co/g0m9cEUz0T%3A80%2Fannounce
RELEASE a6bbd9affe0c2725c1b7410d66833e24
Noah Santoni retweeted
Great content as always from @MatthewBerman!
AutoGen with Ollama/LiteLLM - Each Agent With Its OWN LOCAL MODEL (Tutor... youtu.be/y7wMTwJN7rA?si=dM7f…
Noah Santoni retweeted
Install Stable Video on Your Laptop with 1 Click.
Introducing a brand-new 1 Click Installer for ComfyUI, featuring full Stable Video Diffusion support!
Generate video on your LOCAL machine with Stable Video and run it for free, unlimited, with one click.
And it's REALLY fast!
Noah Santoni retweeted
We're rolling out new features and improvements that developers have been asking for:
1. Our new model GPT-4 Turbo supports 128K context and has fresher knowledge than GPT-4. Its input and output tokens are respectively 3× and 2× less expensive than GPT-4. It’s available now to all developers in preview.
2. Assistants API and new tools (Retrieval, Code Interpreter) will help developers build world-class AI assistants within their own apps.
3. The platform is becoming multimodal. GPT-4 Turbo with Vision, DALL·E 3, and text-to-speech are all now available to developers.
Oh… and we’re doubling GPT-4 rate limits. openai.com/blog/new-models-a…
Noah Santoni retweeted
Introducing pplx-api, our LLM API which serves Mistral and Llama2 models with blazing speed and throughput. pplx-api is in public beta for our Pro subscribers!
We partnered with @nvidia and @awscloud to build our proprietary inference. Learn more: pplx.ai/introducing-pplx-api
Noah Santoni retweeted
Science Paper, "A principal odor map unifies diverse tasks in olfactory perception"
science.org/doi/10.1126/scie…
Open access PDF: biorxiv.org/content/10.1101/…
Noah Santoni retweeted
Introducing TinyLlama🦙
Undoubtedly one of the most impressive LLM projects I've seen recently!
-- 🌟 Introduction --
The ambitious TinyLlama project aims to pretrain a 1.1B Llama model on 3 trillion tokens – yes, 3 trillion! 🚀
The idea is to achieve this goal within "just" 90 days using 16 A100-40G GPUs 🚀🚀 with proper optimization.
The training commenced on 2023-09-01, and you can witness its convergence live (find the link below).
-- 🏛️ Architecture --
They've adopted the exact architecture and tokenizer as Llama 2. This means TinyLlama can seamlessly integrate into numerous open-source projects built upon Llama.
Furthermore, TinyLlama is compact, with only 1.1B parameters. This compact design enables it to cater to various applications that require limited computational and memory resources.
-- 📚 Dataset --
The dataset consists of 3 trillion tokens sampled from a mix of 70% Slimpajama and 30% Starcoderdata, with the GitHub subset of Slimpajama excluded.
-- ⚙️ Optimizations --
They're utilizing a variety of fancy optimization techniques such as flash attention 2, fused layernorm, fused swiglu, fused cross-entropy loss and fused rotary positional embedding.
These optimizations result in a throughput of 24k tokens per second per A100-40G GPU, achieving 56% model FLOPs utilization (an incredible feat! 💯).
This has the potential to be a game-changer for end devices, as the 4-bit-quantized TinyLlama-1.1B's weight only occupies 550MB of RAM! 💥
-- 📊 Status --
They'll release intermediate checkpoints according to the schedule provided below. The first released checkpoint is already in competition with StableLM-Alpha-3B & Pythia-1B.
-- 🔗 Links --
In the meantime, you can track the live cross-entropy loss via this link: wandb.ai/lance777/lightning_…
Find the repository here: github.com/jzhang38/TinyLlam…
Noah Santoni retweeted
I was chatting on Discord with an anon account ,who had just come up with a really smart new approach to model merging and had some encouraging results to show.
But then he had to go.
Turns out he's in 10th grade and was in trouble because he wasn't focussing on English class.
Noah Santoni retweeted
“DoctorGPT is an LLM that can pass the US Medical Licensing Exam. It works offline, it's cross-platform, & your health data stays private”
Working on a new local model of this that will use telemetry from 16 human sensors to create a daily health check. github.com/llSourcell/Doctor…

























