April: AI Highlights & More

We might be in a bear market, but invention is at a feverish, new peak (the last time i felt this level of excitement was ~2017). ICYMI, these are just some of the new capabilities that emerged over the last 2 weeks. 🤯

1. AI Agents Are Here

The idea of AI agents (autonomous agents that act on your behalf) are now truly a (usable) reality. AutoGPT allows GPT-4 to prompt itself, which means you can give it a goal, and AutoGPT will come up with a list of tasks and then execute them—all by itself. A great thread on use-cases and how to get it set up here: https://lnkd.in/g44QXkVY

2. The barriers to software development have gone down to zero

Now you can ask GPT-4 to generate code from scratch, without even knowing how to code. Designer and entrepreneur Morten Just asked GPT-4 to create an app for movie recommendations and now it’s in the app store. In a historic moment on Twitter, VC and AI tinkerer Yohei Nakajima, who is not a coder, created one of the most popular AI agent experiments (named “BabyAGI”) using GPT-4.

Links: https://lnkd.in/g-gFpk97

3. AI meets the web

LLMs get access to the internet (the data for present-day models backdate to 2021). This is a paradigm shift not just how in how we surf the web, but how we interact with the world. Your personal AI agent can book flights, research topics, do tasks — all right from your browser. Link: https://lnkd.in/gDm9M2cb

4. Ethereum completes migration to proof-of-stake

One of the world’s largest open-source communities, Ethereum, finalizes its multi-year migration to a proof-of-stake network, which is a model that is 99% more energy efficient and much more scalable. The hard fork is a milestone that completes phase 1 of Ethereum 2.0 (called ‘The Merge’), and precedes phase 2 (The Surge), which will implement sharding to improve latency, performance, and scalability. Link: https://lnkd.in/gJGFmHGW