Cristian

🔥 - Selfhosted Web Analytics
🥇 - AI Benchmarks
🎹 - Free piano learning app

15+ years SWE experience.

Love 🏓

[Cristian @XCSme]

I have to say this again, and again, but it is VERY IMPORTANT when working with LLMs and making sure you don't run out of tokens.

Do not ask the AI to do something only once (i.e. verify a screenshot), ask the AI to create a system/script to always automatically verify and report those issues in an optimized way.

Also, once a system is implemented, don't stop there. Ask it again, after a few runs, to check the executions and see if it can improve performance/efficiency of the system. And, not only once, this improvement check should be done regularly over time, as your tasks or data do change, so the system has to be "recalibrated".

For example, I just tried using Cludflare crawl/Browser Run to get some info automatically, and Astra xhigh implemented the basic API call to get the data. After this, I suggested to try again, but see if it can reduce the Browser Run time or load, while still retrieving the same data, it did it:

"The light web retrieval test worked: 0.78 browser seconds, down from 4.34 seconds—about 82% less for the same product. Data accuracy checks still passed."

[Cristian @XCSme]

@miguelgbandeira Oh, seems to be back up. Probably just an outage

[Cristian @XCSme]

Anyone knows any https://t.co/JkRgr4HG4g (former https://t.co/WzJlrifJME) alternatives?

It seems to be gone.

Websites where I can just CTRL+V images from clipboard to share them.

[Cristian @XCSme]

Am I dumb? What even is this question from Opus:

"Should the vendored flipMath.js ever be re-extracted from the patched page-flip package, or can the page-flip dependency and its patch be dropped once the orphan PlacementPageTurn.tsx is removed?"

[Cristian @XCSme]

Oh ok, so rendering is at 120fps and new data is streamed from the server. Makes sense.

It does say ~90FPS in top left from what I see though, but looks good!

You can ask it to see if can reduce the number of draw calls (actually, what GPU API does it even use for rendering?)

[Cristian @XCSme]

@mosesbabychrist But why would the streaming be done at 120fps, and not in chunks, once a second for example?

[Cristian @XCSme]

@Mai97co @thsottiaux Yes!! Pleeease tibo, fix the buggy/slow app first before adding anymore features, it's almost unusable :(

[Cristian @XCSme]

@kevinace @bryan_johnson So they are just points I guess...

[Cristian @XCSme]

@victormustar Nice demo!

At some point I tried something similar with a neural network, (predict mouse movement) but that didn't work but because neural nets don't really like having two different outputs for the same input.

[Cristian @XCSme]

Yeah, it's very weird, and the resets make it almost impossible to track.

Unless they release a clearly better SOTA model within 2 weeks, I will cancel my sub, Claude UI is better, less buggy, better value for money, and feels nicer to chat to.

I love Astra for just doing things, without asking too many questions, but $200 per month just to get constant errors and random usage drops makes no sense.

[Cristian @XCSme]

@thsottiaux That's a great feature!

But will it properly work? Now even sending basic prompts or steers fails 50% of the time via ChatGPT remote

[Cristian @XCSme]

Wait, with Claude plans you also get $100 API credits for free per month, on top of your weekly usage??

[Cristian @XCSme]

@ProtonDrive I usually validate it by having my password manager show saved logins for that domain, if it doesn't show them, it means domain is fake

[Cristian @XCSme]

@elonmusk @IterIntellectus Wouldn't that lead to inflation?

[Cristian @XCSme]

@elonmusk @SmokeAwayyy 1984 vibes

[Cristian @XCSme]

@ManifestAverage @bryan_johnson It's over...

[Cristian @XCSme]

That Retry timer is quite new, never saw it before.

[Cristian @XCSme]

I am almost done with Codex, I switched from Astra to Sol 6.1 Fast, so I can complete something faster, and it asks me to wait 3 minutes before I can continue, lol

[Cristian @XCSme]

@JackEllis Companies like this, i.e. intercom, sometimes charge per conversation (for their AI support for example), so don't have many incentives to properly filter all spam automatically I guess.

[Cristian @XCSme]

POV: you post your new app on Reddit and start reading all the positive and constructive feedback.