zlacker

>>swolpe+(OP)
When even Simon falls for the hype, you know the entire field is a bubble. And I say that as an AI researcher with papers on LLMs and several apps built around them.

Seriously, until when are people going to re-invent the wheel and claim it's "the next best thing"?

n8n already did what OpenClaw does. And anyone using Steipete's software already knows how fragile and bs his code is. The fact that Codexbar (also by Steipete) takes 7GB of RAM on macOS shows just how little attention to performance/design he pays to his apps.

I'm sick and tired of this vicious cycle; X invents Y at month Z, then X' re-invents it and calls it Y' at month Z' where Z' - Z ≤ 12mo.

>>behnam+wa
Not disagreeing with anything you said except:

> The fact that Codexbar (also by Steipete) takes 7GB of RAM on macOS shows just how little attention to performance/design he pays to his apps.

It's been running for weeks on my laptop and it's using 210MB of ram currently. Now, the quality _is_ not great and I get prompted at least once a day to enter my keychain access so I'm going to uninstall it (I've just been procrastinating).

>>joshst+4b
Last I checked it spawns claude subprocesses that quickly eat up your RAM and CPU cycles. When I realized the UI redraws are blocking (!) I noped out of it.