Lars

[Lars @0xreself]

@tomhaerter @samlambert Higher TPS is nice, but when you look at the average cost per million tokens. Is it competitive compared to the API?

[Lars @0xreself]

@tomhaerter @samlambert is that even worth price wise?

[Lars @0xreself]

Does anybody ever use Godot to create synthetic data for AI training?

I actually think it's not too bad, even though the headless mode can't render stuff.

[Lars @0xreself]

@rand_longevity that message is a sign of low IQ

[Lars @0xreself]

@nia_thinks This is not an issue about thinking clearly; the issue is the ambiguity of the English language or language in general.

So the LLM interprets what you could mean.

[Lars @0xreself]

The approach from @ThePrimeagen with 99 Nvim might be a way to achieve more of what mentioned.

Has someone tried it out already?

[Lars @0xreself]

In general, I believe that if you have technical knowledge of what is going on with your codebase, you will be better at leveraging AI.

You should make decisions; the AI should just implement them.

[Lars @0xreself]

For sure AI is only good. I mean its AI...

[Rohan Paul @rohanpaul_ai]

Anthropic's own study proves Vibe-Coding and AI coding assistants harm skill building.

"AI use impairs conceptual understanding, code reading, and debugging abilities, without delivering significant efficiency gains on average"

Developers learning 1 new Python library scored 17% lower on tests when using AI.

Delegating code generation to AI stops you from actually understanding the software.

Using AI did not make the programmers statistically faster at completing tasks.

Participants wasted time writing prompts instead of actually coding.

Scores crashed below 40% when developers let AI write everything.

Developers who only asked AI for simple concepts scored above 65%.

Managers should not pressure engineers to use AI for endless productivity.

Forcing top speed means workers lose the ability to debug systems later.

----

Paper Link – arxiv. org/abs/2601.20245

Paper Title: "How AI Impacts Skill Formation"

[Lars @0xreself]

@cop_on_fire @rohanpaul_ai I argue that in some way knowing function signatures correlates with knowing the codebase.

And knowing the codebase helps with understanding the overall system, which leads to better decision making later on.

[Lars @0xreself]

@tomhaerter Same issue its a bit more ergonomic then latex. I am talking more about application in what you write in then the tech that is powering it.

And tbh latex and typst have similar structure in how you use them, just latex feels a bit more before 2000.

[Lars @0xreself]

Why isnt there a good latex writing program with ai integration?

I am not a good writer, so formulating my thoughts in a paragraph and let a LLM rewrite it to be fluent and correct would be so great.

Current workflow is copy pasting from LLM to overleaf and back...

[Lars @0xreself]

@tomhaerter if you have a spare B200 I would take one

[Lars @0xreself]

just everything is bloated nowadays...

Things that are good should just stay like they are we don't need some weird as hell feature that 1% of the user base is using.

[Peter Yang @petergyang]

Why does reading some PDFs require 2 GB of space?

Even Adobe's software is bloated.

[Lars @0xreself]

@CodeByNZ - Build your own bullet point list that does not have wrong newlines

[Lars @0xreself]

Making something intelligent should not be about scaling compute to insane amounts... that is more brute-forcing.

It's about making the core system better.

[Lars @0xreself]

needing 5M tokens to solve a Sudoku is crazy.

There is something called SAT solver that do not need a full data center to run.

[Christos Tzamos @ChristosTzamos]

1/4 LLMs solve research grade math problems but struggle with basic calculations. We bridge this gap by turning them to computers.

We built a computer INSIDE a transformer that can run programs for millions of steps in seconds solving even the hardest Sudokus with 100% accuracy

[Lars @0xreself]

@robjama contact @aaronmahlke if freelancing works for you great stuff he does

[Lars @0xreself]

@DylanKelly_DK ...wait so knowing nothing is good?

[Lars @0xreself]

@echo_vick and paying for it👀