Martin Bel

...

[Martin Bel @__mbel__]

@rasbt Actually the opposite works pretty well. Larger models with low or medium effort are pretty good for planning/research.

[Matt Pocock @mattpocockuk]

The "should you read code" debate is dumb because the real decision isn't binary, it's a scale:

1. Reading every line of every diff
2. Scanning every diff, reviewing important lines
3. Ignoring diffs but understanding the 'why' of every PR
4. Spot checking PR's instead of reading every one
5. Ignoring PR's, but doing regular spot checks on the codebase
6. Ignoring the code, but spot checking agent traces to help improve the system
7. Ignoring both the code and the system, let models handle everything

Where are you on the scale?

[ali @waterloo_intern]

try the model yourself here
https://t.co/B1cIvnkRug

[Martin Bel @__mbel__]

@beyang Useful for constants for example

[Martin Bel @__mbel__]

@francoisfleuret why not going with 8b directly. 1b is just too small.

[Martin Bel @__mbel__]

@ansonlin Can you discount their future earnings?

[Martin Bel @__mbel__]

@FernandoMarull Esperar a las elecciones para que sea "útil" la noticia

[Martin Bel @__mbel__]

@badlogicgames @lutfi_gr Torch was built on Lua initially

[Martin Bel @__mbel__]

@FernandoMarull @fede_machado_b Entro overplotting al chat

[Martin Bel @__mbel__]

@badlogicgames Is the extension public? Looks interesting!

[Wes McKinney @wesmckinn]

Just released my next-generation agent session viewer with analytics dashboard (Go + Svelte). A much evolved version of wesm/agent-session-viewer which is now deprecated:

https://t.co/D4BDX5cfX7

[Martin Bel @__mbel__]

@badlogicgames What are you building it for?

[Daniel Howells @howells]

I made @figma, but everything is text. Create a quick wireframe, export as markdown, and chuck it into @claudeai. Try it out at https://t.co/G8hdmuQQbJ

[Guillermo Rauch @rauchg]

https://t.co/U8rutTJffK

[Martin Bel @__mbel__]

@alerosenfeld suena a fruta

[Martin Bel @__mbel__]

@lucasllach empezá de 1970... Le pega un paseo épico al S&P 500

[Martin Bel @__mbel__]

@dexhorthy They are annoying

[Martin Bel @__mbel__]

@ahmetb Most likely it was trained to do this well during RL.
It seems this can be a verifiable problem.

[Martin Bel @__mbel__]

@intellectronica @zeddotdev Are you managing one venv for each skill and calling it with uv?

[Martin Bel @__mbel__]

@_jaydeepkarale stocks in argentina are fun, here is an example!

~65 - 2017
~30 - 2018
~8 - 2021 ( populist government killed the market)
~50 - now