@adlrocha - Running LLMs and ML in Wasm

Searching new runtimes for AI

🗞️ Hot Off the Press

• [Blog] Running LLMs and ML in Wasm: The outcome of my research about running LLMs and ML models in Wasm and heterogeneous runtimes.

• [TIL] Enabling Hibernate on my Linux laptop (Pop OS): I was sick of closing my laptop and coming back to a device without battery. This TIL walks you through how I configured hibernation on my Linux distro to prevent that from happening

  

🍙 Byte-sized bites

• [👨🏻‍💻] Browser extensions are underrated: the promise of hackable software: Oldie but goodie. Extensibility of the web. The need to hack computation on our surroundings (starting from the web).

• [⛓️] Read, write, own manifesto from a16z: Interesting read about the impact that blockchains can have for the Internet.

• [🚀, 🔬, 🤖, 🖧] Beff Jezos presentation on e/acc and the need for decentralization. If you don’t know what e/acc and who Jeff Bezos are, this Lex Fridman conversation is definitely a must for you.

• [👨🏻‍💻] Decisions I endorse and regret after 4 years running infrastructure at a startup: If you had run any kind of infrastructure throughout your career, you may have faced many of these decisions yourself. Definitely a great “go-to” to reinforce your opinions about the different options when faced with these decisions again.

• [👨🏻‍💻] Is Mojo faster than Rust?: An interesting (and obviously biased) blog post about Mojo, a new AI-oriented programming language aiming to replace Python. If you are curious about what Mojo is, I learnt about this (again) in this Lex Fridman conversation (definitely worth a listen). This post is also slightly related with my new blog post linked above.

• [💰] YC Request for Startups: Looking for an “investable” idea? YCombinator has already released their request for startups.

• [💾] AMD Quietly Funded A Drop-In CUDA Implementation Built On ROCm: It's Now Open-Source