NOTE: This was another light week where I didn’t get to read up on as many posts as I wanted so not too many links today.

Qwen 2.5 VL!
By Simon Willison

I haven’t played with this model yet, but looking forward to another multimodal model I can run locally on my home server.

Aider

This is something I’m adding to my toolkit to try out in the near future. I’ve seen people mention how great it is online, but I haven’t actually tried it for myself since I have access to other cloud hosted AI tooling. I’m interested to see how this works with my locally hosted models though and to see how this works for me given I’m usually using an IDE for dev work.

Cursor Project Rules (via X)
By Michael Feldstein

Interesting reading through some of these project rules to help Cursor be more helpful without having to put a lot of content directly in the prompt. This seems more flexible than shoving everything into the .cursorrules file while still letting Cursor see whatever context it might need.

VGHF Digital Archive

This isn’t AI or even work related, but if you’re around my same age and were into video games growing up, you might find this interesting. It’s scans of all kinds of popular gaming magazines from the 90s/2000s. I used to *love* reading so many of these. It’s awesome to see them available on the web.

NVIDIA Puts Grace Blackwell on Every Desk and at Every AI Developer’s Fingertips

I don’t know if I’ll personally try to buy one of these or not, but I did sign up for the email list to see when it’s available. This sounds like a great idea for anyone wanting to do experiments locally with larger AI models. I’m running some larger models on a 96GB of RAM server, but it’s all on the CPU so it’s super slow. This Digits device seems relatively cost effective (for what it is). It sounds like a 5090 might be more powerful, but way less RAM, so this is a good balance of GPU power with unified RAM. Yes, you can get an Apple machine that can match the unified RAM, but at a higher cost AND I’ll be very, very shocked if this machine’s performance doesn’t blow Apple Silicon out of the water for these AI model specific tasks (it may be worse in every other way, but not that one).

34 Life Lessons from 34 Years The Curiosity Chronicle
By Sahil Bloom

Lots of good advice in this one. I always enjoy seeing what general life lessons learned posts. I can certainly relate to quite a few of them, but I always find a few things that make a lot of sense, but I’ve never really thought about before.

Agents
By Simon Willison

This is Simon’s commentary on another article. You should read the original article too. I’m just getting started understanding agents. I certainly can’t process everything in the longer article, but I’m going to come back to this once I understand some basics better.

I Quit! The Tsunami of Burnout Few See
By Charles Smith

I think this is an interesting take on why we see so much burnout in the tech industry (I’m sure it’s everywhere, but I’m most familiar with tech). I think there are quite a few good points in here about what’s happening and how it’s not just “overwork” that causes all of this – it’s larger than that.

Mastering Long Codebases with Cursor, Gemini, and Claude A Practical Guide
By wheattoast11

I haven’t tried this yet, but I do a good bit of work with Cursor, so I’m really interested in giving this a try for some larger codebases I work with. It sounds like a great idea and I can see how it would help. I’m hopefully going to get a chance to try this out in the next couple of weeks.

Okay, Gemini 2.0 might actually be better than ChatGPT Pro. (Twitter Thread)
By Leo Grundstrom

This is a Twitter thread, but I thought it was interesting enough to include here. It’s a handful of use cases where Gemini 2.0 beats ChatGPT Pro even though Gemini is free (at least free enough for smaller experiments). I don’t work with Gemini much, but I have wired it into my home Open WebUI server so I’m going to give some of these a spin.