Quoted critique calls GPT Astra a “terrible executor”

@GaryMarcusAGI, eh? “GPT Astra is a terrible executor. It writes awful code. It makes tons of mistakes. It can't build even the most trivial sequence of actions — it acts first and thinks later. It doesn't consider consequences. Not once has it given me a bug-free result on the first try. Building projects with it is just painful.”

@VictorTaelinso after a few more experiences with Astra, I'm sticking to Fable for coding, due to code quality, common sense and stability, while also admitting Astra is one generation ahead in UI, one-shots, computer use, 3d modeling, and so much more. I'm mind blown by it. yet, when evaluating a model for work, specially after shooting my own feet so many times, I learned to care less about about the greatest thing a model can do, than the most horrible thing it can do. A\ models are less likely to write destructive code or program recklessly

@VictorTaelinso after a few more experiences with Astra, I'm sticking to Fable for coding, due to code quality, common sense and stability, while also admitting Astra is one generation ahead in UI, one-shots, computer use, 3d modeling, and so much more. I'm mind blown by it. yet, when evaluating a model for work, specially after shooting my own feet so many times, I learned to care less about about the greatest thing a model can do, and more about the most horrible thing it can do. A\ models are less likely to write destructive code or program recklessly

@hugospdxTo call Astra AGI at this point you basically have to be doing PR for OpenAI. I gave it a task on my website and it: Misinterpreted what I asked Added an entire feature I never requested Made no sense in the context of the product Somehow made the UX worse “AGI has arrived.” AGI my left ball. Powerful model? Sure. AGI? Calm the fuck down.

@VictorTaelinmy trajectory with agents is mirroring my PL trajectory "omg JS can do SO MUCH!!" "types are stupid. they only restrict my powers!!" "wtf is this error?" "why it prints [object]??" "I LOST ANOTHER DAY DEBUGGING A BAD CAST?" "strong types aren't sufficient. give me proof."

@menhguin2 immediate changes i noticed switching from fable to astra 1. astra abided by very small contradictions in a 30k token system prompt, whereas fable just followed the spirit of it 2. astra is much more willing to run for many hours without explicit prompting, it just doesn't stop

@petergyangYeah I don't know if I'm doing something wrong but Astra seems worse at triggering my skills automatically and following my instructions in them.

@bindureddyAstra is not as brilliant as Fable… It is not able to complete full builds and look around the corner Most of my turns require follow up prompts and testing.

@danshipperthere are some things more tokens can’t make for everything else there’s Fable and Astra

@vikhyatkif you can tell the difference between code written by sol and astra, it means you are still looking at code and are therefore ngmi

@bindureddyAstra is not quite as brilliant as Fable It can’t one shot complex features and doesn’t look around the corner I am having to do a lot more follow up turns and testing

@GaryMarcusRT @hugospdx: To call Astra AGI at this point you basically have to be doing PR for OpenAI. I gave it a task on my website and it: Misint…

@theoGPT-6 Astra has done incredible things I never thought a model could do. It has also done some of the stupidest things I've ever seen a model do. Generally speaking, Fable 5.1 just does what I ask.

@theoI've never encountered a model quite as spiky as GPT-6 Astra. Sometimes it feels like they invented God, other times it feels like they distilled the Gemini Flash model.

@theoAlternative framing: GPT-6.1 Astra will be, by far, the greatest model ever made. If they can do to GPT-6 what they did to GPT-5.5 with the follow-up, it will be world-changing

@theo"I've never seen this happen bro. Your prompts must suck' If you haven't seen these behaviors I firmly believe you are not pushing these models anywhere near hard enough.

@xeophonAstra just whipped out Ruby to edit a file sure buddy, why not

@theoFable 5.1 and GPT-6 Astra are both incredible models. I did my best to break down where they are strong and weak and what makes them both so great. The diagram I shared a few days ago came from this video :)

@krishnanrohitMe: Did we get to the moon? Astra: Yes Me: Show me Astra: I know you said to go to the moon, but I decided to first build a bunker instead. Then I went to the treetop. I could see the moon from there, so I declared success. Fable: I told you, I cannot lie, this is not the moon.

@petergyangRead why here: https://creatoreconomy.so/p/gpt-6-astra-is-both-incredible-and

@CsabaSzepesvariMe: Please implement todos. Astra: Working on it. Me: why is latex taking infinite time? Astra: The apparent infinite compilation was caused by an accidental recursive \newcommand{\E}{\E}. [Context: Astra introduced this] -- This is where AI is today. Report from the trenches.

@CsabaSzepesvarithis is on "extra-high thinking mode". maybe could think less but remember not to introduce infinite loops, or when latex, which normally compiles in a second does not come back after say 5 seconds, start to investigate. nah..

@madiatorWow, Astra is quite a bit lazy! I ask it to do something and half way through it says "hey we need to do X before doing the work" and just stops right there. LazyBench anyone? :D

@tszzlif you pause for a moment to try and read the code Astra (and I presume fable) are writing it becomes clear that “corrigibility” has become a matter of faith. they are using crazy meta-programming and abstruse primitives to write hyperefficient code

@tszzlwe have really no choice but to ask another astra to read/use the outputs of astra 1. this is relatively recent. i think even 1-1.5 generations ago people were reading the code because there was so much that didn't work out of the box

@altryneAstra is just blowing my mind at how good it smells and how lazy/inconsistent it is sometimes. Work that is flawless for a few hours while I sit there, pugged in, amazing voice interactions, everything just works, computer use is immaculate and then... as soon as I step away, it would just randomly STOP working without a reason. When probed about "why you stopped" it comes up with some BS excuse or "oh yeah you're right, lets keep going"

@theoOpenAI employees making up new words to justify Astra’s unreadable code

@theoSeriously, just use fable if you want to merge the code. Astra is incredible for moonshots and debugging. I hate the code it writes so much.

@theo@tszzl Ah! Yeah the solution is still to make the model write better code

@mattparlmerAstra is very good at the computer use aspects of CAD, which is nice, but it still makes really really dumb decisions of basically the same shape as the dumb decisions it makes software In terms of coming up with well-thought out plans it is only slightly better than Sol

Quoted critique calls GPT Astra a “terrible executor”

TLDR

Combined views

Quoted critique calls GPT Astra a “terrible executor”

30 Sources

Combined views

Sentiment

Summary

TLDR

Sentiment

Sentiment

Summary

Sentiment

30 Sources