Rishabh Bhargava

Excited to announce that Refuel is officially out of stealth today! We’re building a platform to create, clean, enrich and label datasets at the speed of thought, using LLMs. 🚀 More details below.

@teknium We have a really fast Whisper API (that scales well) at @togethercompute : together.ai/blog/speech-to… Can help you get started if you'd like.

Does whisper have an api somewhere or is there better transcriber as a service models

@viccpoes would be amazing!

should we open-source?

@tnm What are you doing with tides data, btw?

@rish_bhargava data for tides, sea conditions, wind, weather, & cloud prices

By far the biggest issue with MCP is not the protocol itself (it's fine, it's just a protocol lol), but the total shit state of actual MCP servers out there. A few things right now, both technical and social: a) You still have to think about actually designing these things, you can't just throw a bunch of tools out and assume it'll work b) Resources and tools are not the same thing. (clients need to catch-up to this ofc) c) MCP is absolutely not REST. Don't map REST to this. d) Threadlad overstatement of effectiveness of existing servers is causing massive cognitive dissonance among users when they try stuff out e) Accuracy of LLM tool use still matters and is not perfect (or even great!)

@tnm I wonder if @rejuvyesh and @SilurianAI team can get us something great for weather data. Cloud prices is a big pain for sure 👀

@adityagrover_ Congrats @adityagrover_ ! Well deserved

Thank you for the honor — truly an acknowledgment of the tireless efforts of all my students, mentors, collaborators, friends and family over the years!

Congratulations to our portfolio companies, @togethercompute and @RefuelAI, on uniting their strengths to power the next generation of AI infrastructure! Together AI’s AI Acceleration Cloud enables developers and enterprises to train and deploy generative AI models with speed, control, and cost-efficiency. By bringing in Refuel’s purpose-built data models and orchestration platform, they’re tackling one of the biggest bottlenecks in AI: cleaning and structuring messy data at scale. We’re proud of @vipulved, @rish_bhargava, @nihit_desai, and both teams for building the foundation that enables applied AI to thrive.

We have some big news to share today - @RefuelAI is joining @togethercompute to help accelerate the future of open source and enterprise AI! together.ai/blog/together-…

@auren We'd seen AI models outperforming humans even with earlier, less powerful LLMs - see this report we published in June 2023: refuel.ai/blog-posts/llm… We didn't measure performance of humans in the first few minutes, but makes sense that humans would quickly get bored here!

we ran a contest: Humans v AI on data labeling. had complex data labeling rote task. initial finding: humans beat AI but ... humans only beat AI for the first 3 minutes. then humans got bored. an hour in, humans were terrible. AI was consistent over time. so: AI beat humans

@adityagrover_ Congrats @adityagrover_ !

A few months ago, we started Inception Labs, a new generative AI startup with a rockstar founding team. At Inception, we are challenging the status quo for language generation. Our first results bring blazing fast speeds at 1000+ tokens/sec while matching the quality of leading speed-optimized frontier LLMs. And all on commodity NVIDIA H100s - an industry first! Our vision is to extend the frontier of speed, quality, and cost for next-generation language models. Join us!

To Reason or Not to Reason, that is the question 🤔

@natfriedman Will this continue until tomorrow by any chance?

Need volunteers to come to my office in Palo Alto today to construct a 5000 piece Lego set. Will provide pizza. Have to sign NDA. Please DM

@martin_casado yes please!

Hey infra folks. We're standing up a new Discord server to discuss CS infra. If you want an invite DM me (reply and I'll follow). thanks!

Data intelligence too cheap to meter RefuelLLM-2-mini (75.02%), our latest 1.5B param SLM, outperforms all comparable models including Phi-3.5 (65.3%), Qwen2.5 (67.62%), Gemma2 (64.52%), Llama3-3B (55.8%) and Llama3-1B (39.92%) across our benchmark of data processing tasks such as labeling, enrichment and structure extraction RefuelLLM-2-mini is a Qwen2-1.5B base model, trained on a corpus of 2750+ datasets spanning tasks such as classification, reading comprehension, structured attribute extraction and entity resolution, using the same recipe as other models in the Refuel-LLM family. It's fast! We’re open sourcing the model weights, available on @huggingface - huggingface.co/refuelai/Qwen-… If you'd like to access models, along with fine tuning support, DM me or reach out to us: refuel.ai/get-started Grateful to our early customers for their partnership, and the entire @RefuelAI team for their hard work 🚀

@patio11 This is an exceptional piece of writing!

Piece is here, incidentally: bitsaboutmoney.com/archive/debank…

I am (sincerely!) pleased how many crypto enthusiasts have enjoyed the piece on debanking.

@davj @kendallamesphil Congrats!

Grateful to welcome our baby girl into the world She’s our dream come true ❤️

@LeopolisDream we should chat @RefuelAI :)

The most boring task you might have to deal with is data cleaning. Also record deduplication. And there is no winning solution to this. Only manual tedious work. Should I try LLM to help with this or this will add another layer of issues?

At the summit, @tomas_hk led a panel featuring @rish_bhargava, @corbtt, and @JunchenJiang discussing "Scaling AI Operations for Diverse Workloads." #GenAISummit

@ali_k_chaudhry Have you tried @RefuelAI - pretty reliable from what I've seen 😉

@vishakhhegde @Sierra_Ventures @ycombinator @Soma_Capital @generalcatalyst Congrats @vishakhhegde and team!!

🚀 Excited to share that Inventive AI has raised $4M Seed led by @Sierra_Ventures and participated in by @ycombinator, @Soma_Capital, @generalcatalyst for our mission to build the world's best sales intelligence platform. We are building for AI-first RFP (Request for Proposal) and sales questionnaire workflows that cut busywork by over 70%. This will allow sales team to focus on what really matters - winning the proposal! 💪 Special thanks to our customers, investors and talented teammates who believe in our vision. We’re pumped to push the boundaries of what’s possible in sales intelligence! Check out the press release and some of the news articles in the first comment. @dhiren9, @gaurav_nemade15 and I are super grateful to our investors Mark Fernandes, Anne Gherini, @dessaigne, @koomen, @aaron_epstein, @daltonc, @aneelr, Nikhita Jaaswal & Tamanna Khemani.