Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Astra for Coding: Why Are We Doing This Again?

lucumr.pocoo.org

Post reply on HN
Print view View original

301–310 of 334 posts

Re: Astra for Coding: Why Are We Doing This Again?

#301

Post by huurtehoog »

tingletech wrote: post #300
huurtehoog wrote: post #294

Earlier quoted context omitted.

Reading Snowcrash, the guy gave up coding because "back in the day" a hacker could do it all alone but now it was all about software factories. That was in a book written in 2000. Software is theory building, as Naur puts it. It's a learning process, a research project. Orgs have been trying to turn it into assembly line work forever. There's a lot of money in it. I don't care. It has also never been easier to solo h…

Neal Stephenson's Snow Crash came out in 1992. I guess the paper back became popular in 2000 though.

Thanks I was going by the date I glanced at the counter cover

Re: Astra for Coding: Why Are We Doing This Again?

#302

Post by generic92034 »

bob1029 wrote: post #148
generic92034 wrote: post #54

Earlier quoted context omitted.

> LLMs will never be able That is a very bold claim, unless you meant "current LLMs".

The coordinate system in game scenes is known as "world space", but that rarely means our world. How do you train an LLM to create a world that only exists in an artist's head? I think spending a day with just the lighting systems alone would alleviate us of any misunderstandings here. Getting lighting to work right isn't something you can solve by duct taping a vision model to the contraption.

Again, my issue with the statement is not about current LLMs. Who knows if LLMs cannot be a major component of AGI, or even ASI systems. Then the capability to do whatever humans can do is part of the definition.

Re: Astra for Coding: Why Are We Doing This Again?

#303

Post by generic92034 »

rhdunn wrote: post #135
generic92034 wrote: post #54

Earlier quoted context omitted.

> LLMs will never be able That is a very bold claim, unless you meant "current LLMs".

These models aren't really LLMs -- they don't just operate on text tokens. They often include vision models and in some cases audio models. That means that they can better associate the meaning of images and words together so that when someone says "make this button blue" or "create a 3D model of a rocket" they have some level of understanding of what that is and what needs to be done. The key question is how good th…

Again, I was not talking about current LLMs. The statement was made without any limit regarding time. Who knows how far LLMs can get, in the end?

Re: Astra for Coding: Why Are We Doing This Again?

#304

Post by valdork59 »

coldtea wrote: post #127

> I’m more and more convinced that all of AI engineering is Neijuan (内卷, meaning curl inwards). In China it describes a system that demands ever more effort and competition without improving output. The way in which it sometimes shows up in the West is the 996 nonsense. The English term for Neijuan is “Involution” from the book Agricultural Involution. Agricultural involution describes the intensification of farming…

no this term is more complicated while not expressing much more

Re: Astra for Coding: Why Are We Doing This Again?

#306

Post by juancn »

Anything shitty that enters the context window shifts the entire thing to shittiness. So far this has been my experience on pretty much any model.

Context feeds on its output.

Once you it goes that road, unless you stop it and give it enough counter examples and details of what you want (i.e. you're nudging it on latent space towards a better spot), it keeps degenerating.

It gets even worse if the context window is compressed before you get a chance to correct.

Long horizon agents can degenerate at machine speed.

I still think you get much better results if you give them short horizon, well specified tasks.

Re: Astra for Coding: Why Are We Doing This Again?

#307

Post by owebmaster »

djeastm wrote: post #281
codingisfreedom wrote: post #43

I’ve asked Astra to build me an app for a prototype I created quickly using Sonnet. It’s been 2 days and it made no real progress on the actual app. It created docs, scripts, workflows, and it’s doing a bunch of reviewing on every PR. I told it that I just need an MVP. I’m pretty sure an average senior engineer would have finished that task much quicker, and guaranteed with more readable, higher-quality code. Meanwhi…

>I think I’ve easily crossed 100k tokens so far on nothing 100k tokens? Is it just me or is that very low for an app build?

100k output is a good amount. Probably read and cache are way more, to the millions.

Re: Astra for Coding: Why Are We Doing This Again?

#308

Post by Bluestein »

u8080 wrote: post #117
petesergeant wrote: post #8

Ironically I burned out Fable usage early this week because of Astra using it to run inane full codebase reviews over one line changes, so I have been using Astra extensively. We need a word for “potentially highly capable, but in reality an idiot savant” to describe certain models. No, I don’t need you to write a tmux emulator in bash to test your changes bro, just ask me to run the command.

Exactly this. "I need tool objdump but pacman gcc failed because of no sudo password. Let me write compiler, binutils and disassembling framework"

... and meanwhile all this does is burn more tokens faster which is what providers want - they have no incentive to optimize for succinctness, elegance or compactness if they want you to spend more and more tokens.-

Re: Astra for Coding: Why Are We Doing This Again?

#309

Post by samuell »

VulgarExigency wrote: post #284
samuell wrote: post #138

Earlier quoted context omitted.

Yeah, I think as a general direction we need to allow a useful collaboration with the agents on the issue board. This is btw why Epiq was developed, to keep the board as code, git-backed, distributed (via an event log mechanism), and with the ability to replay the board, to see what agents actually did: https://ljtn.github.io/epiq

How does this compare to Beads, if you don't mind me asking? https://github.com/gastownhall/beads

I hope @jolaflow can chime in here eventually, but from a brief look, my impression, besides the fact that Epiq is based on git as opposed to beads where it is optional, is that Epiq seems to be much more optimized for interactive collaboration between the user and the agents.

The graph visualization in beads surely is a neat thing for showing things, but the replay feature in Epiq should provide a similar understanding of what happened.

But again, it seems to me Epiq is the tool that better allow the user to jump right in and collaborate with the agents on the board.

(Again, this is from a brief look, so I could be missing things).

Re: Astra for Coding: Why Are We Doing This Again?

#310

Post by olwmc »

automatic6131 wrote: post #217
codingisfreedom wrote: post #43

I’ve asked Astra to build me an app for a prototype I created quickly using Sonnet. It’s been 2 days and it made no real progress on the actual app. It created docs, scripts, workflows, and it’s doing a bunch of reviewing on every PR. I told it that I just need an MVP. I’m pretty sure an average senior engineer would have finished that task much quicker, and guaranteed with more readable, higher-quality code. Meanwhi…

>Also, I’m really unimpressed by any “one shot” demo that’s out there in the wild. It means nothing for serious software engineering. If a person, or team of people, can build a demo quickly then it's good odds that they can build the real version (though, famously, not a guarantee). However, it turns out that a machine that can spit out 100 demos of whatever can't actually build the real thing. Similarly, a chess en…

This resonated with me. "Developing ideas and artifacts using AI breaks our normal intuitions along many meaningful axes and we've yet to update" is a really clean idea.
Post reply on HN