latemedium

5 ms·

1. 1.▲
   by latemedium 1y ago

   We need to know if big AI labs are explicitly training models to generate SVGs of pelicans on bicycles. I wouldn't put it past them. But it would be pretty wild in they did!
2. 2.▲
   by latemedium 1y ago

   I think part of the reason why just a few people write custom CUDA / triton kernels is that it's really hard to do well. Languages like Mojo aim to make that much easier, and so hopefully more people will be able to write them (an
3. 3.▲
   by latemedium 1y ago

   There's an important difference between Gemini and Claude that I'm not sure how to quantify. I often use shell-connected LLMs (LLMs with a shell tool enabled) to take care of basic CSV munging / file-sorting tasks for me - I
4. 4.▲
   by latemedium 2y ago

   I hacked something together a while back - a hotkey toggles between standard terminal mode and LLM mode. LLM mode interacts with Claude, and has functions / tool calls to run shell commands, python code, web search, clipboard, and a fe
5. 5.▲
   by latemedium 2y ago

   My experience is starkly different. Today I used LLMs to: 1. Write python code for a new type of loss function I was considering 2. Perform lots of annoying CSV munging ("split this CSV into 4 equal parts", "convert paths in
6. 6.▲
   by latemedium 2y ago

   I'm reminded of the metaphor that these models aren't constructed, they're "grown". It rings true in many ways - and in this context they're like organisms that must be studied using traditional scientific tech