I love that Rex’s primatives are exposed through an expressive CLI — everything is super agent friendly. Looking forward to the Linux client so I can run everything in nixos @superlogical ❄️
I really enjoyed this. It's pretty interesting that some of this advice sounds like MBA-isms. Example:
> "Pick unreasonable goals and tell the model that they’re very reasonable: You need your website to load in 5s instead of 10s? Ask for 200ms. Does something seem impossible to the model? Reiterate that it’s very much solvable, it just needs to keep going."
I think if you've worked in the industry for a while (pre-AI), you've seen some flavor of this from your company's leadership
Hi, I'm Felix. I'm an eng lead working on Claude. I love watching people take flight and do things they never thought possible.
We often describe that as "being more ambitious" when using AI, which is tricky advice to follow. I wrote a tutorial:
felixrieseberg.com/how-to-be…
This capacity shortage will be a major problem. It doesn’t matter how capable AI is if we don’t have the compute to reliably run existing services, let alone spin up newer and more ambitious projects
Andy Jassy says AWS capacity for 2027 is "largely sold out" and a "good chunk" of 2028 capacity is also sold, though AWS reserves some capacity for on-demand (@catherineperlo1 / The Information)
(Visit Techmeme dot com for the link and full context!)
It’s really interesting to see opinions on AI from people outside of tech. Not weighing in on if they are right or wrong — they are as valid as any opinion based on lived experiences. One thing is certain though: the future is here, but it’s not evenly distributed.
few shot vibe coded native mac app to transcribe/persist recordings from my tp-7, which is a delightful device
Using the argmax sdk for on-device transcription github.com/argmaxinc/argmax-…
Eternally a fan of @RogueAmoeba. In a world of slop it’s so refreshing to to have software that is the unique combination of feature rich + intuitive
I don’t think remote sandboxes using frontier models and fast local open-source model inference are incompatible visions of the future. Run agents at night on sandboxes for complex tasks/research, and use fast inference for hands-on computer use.
At scale, the little stuff will always add up.
At 250 billion DNS cache entries, one wasted byte costs 250 GB of RAM.
Five Rust optimizations later: 100 TB freed, inserts 43% faster, lookups 19% faster.
We didn't trade speed for space. cfl.re/4gyL9Jj
The String Seed of Thought technique to inject randomness is super interesting
I'd always thought AI was terrible at design, but after reading today's 🤯 post by @anshuc, I realized I was just doing it wrong.
"AI models are capable of amazing creativity, but that creativity gets stifled. LLMs are trained to be next-token predictors: they look at a sequence of text and predict what typically comes next. Great design is exactly the opposite of this. Great design bends the rules and delights users with memorable, unexpected choices."
@anshuc led design and engineering teams at Apple for 12 years. In his words: "Most people only see 1% of AI's creative potential. I want to show you how to tap into the other 99%."
His 8 techniques for breaking out of the 1%:
1. Use seed strings to inject variety
2. Be much more ambitious with your prompts
3. Create positive feedback loops with subagents
4. Use image generation to enrich designs
5. Use video generation
6. Cut out elements that don’t add value
7. Remove AI tells
8. Rewrite copy by hand
Read the post here: lennysnewsletter.com/p/how-t…
P.S. This design was made by AI 👇