We’re excited to share Portex’s first research paper at the intersection of crypto and AI Check out the paper, "Reputation Oracles: Determining Smart Contract Reputability via Transfer Learning", below: research.portexai.com/reputation-ora…

What We Learned Trying to Build a Blockchain Oracle (and Where We’re Going Next) research.portexai.com/what-we-learne…

1/ In 2024, PortexAI tried to compete with oracle providers like Chainlink neck-and-neck. The economics looked great. Oracle operators on Chainlink alone had generated close to 200M USD in revenue since '21 with take rates at times reaching 46% Here's what happened next..🧵

Introducing the PortexAI Datalab, a full-stack data acquisition platform powered by stablecoins. Let the data flow at datalab.portexai.com

In an era where software is free to create and replicate, the are only lasting moats are: Compute, product distribution, and data.

1/ We're excited about the launch of the PortexAI Datalab: the full-stack data acquisition platform for AI builders. This week, we're showcasing one of the most impactful features on the platform, which has already saved an early user 180 developer hours. 🧵

Helix was trained on a specialized dataset with 500 hours of human tasks. It's an exciting architecture because it uses a single set of neural net weights to learn behaviors. If you feed it specialized data, it will learn new tasks. It's beyond impressive.

1/ New PortexAI research studying the massive growth and reach of Hugging Face, which is unquestionably the heartbeat of open source AI today. Its repository of datasets and models point to a bright future for fine-tuning, and also a coming market for proprietary datasets 🧵

AI is selling off because we've reached a plateau of what foundation models can do without specialized tools. It's like we have an amazing operating system but very few apps running on it. The way forward is fine-tuning tools with better data: the next major trend in AI ⬇️

GPT-4b micro is a model trained exclusively on specialized biological data. It was used to reverse cellular aging with a 50x improvement in efficiency relative to previous approaches. A testament to the power of narrow AI + specialized data. Amazing overview by Rowan Cheung:

"That will only happen with the assurance of effective copyright protection mechanisms, a market for fairly negotiated licensing deals, and additional incentives for creativity, ingenious, and innovative human work." We couldn't agree more 😉

Noticing a trend? Specialized models continue to beat foundation models on task performance, cost, and latency. The emerging design pattern for agents is a foundation-model-brain that can invoke the most optimal tool for a given task.

AI has kicked off a gold rush for data, with OpenAI alone projecting $8B in data-related expenses by 2030. The challenge now is finding a reliable way to value data in this era. Our latest on data valuation techniques: research.portexai.com/data-valuation…

1/ AI isn't just a compute race anymore. It's a data race too. Labs are paying top dollar for differentiated, high-signal data. It's clear now is the time to experiment with new approaches to valuing and incentivizing the creation of frontier AI data. x.com/LucasNuzzi/sta…

Day 2 of #PyTorchCon 🔥 What a ride. Talked with folks using #PyTorch to fine-tune models for drug discovery, cancer research, autonomous vehicles and, of course, customer support! Thanks PyTorch for having us!