aabdi

7 ms·

1.
by aabdi 4d ago
You’re complicating things. There’s no reward for prosocial in llm rl as compared to other targets. Humans have it since prosocial and others have evolutionary reward signals that do.
2.
by aabdi 18d ago
the problem is you have to make the AI watch the whole thing to make sure it works. I've done this sort of with comfyui/same agent factory stuff, but the verification loop only works for models like fable as planner/writer, w
3.
by aabdi 22d ago
I think its hard to understand for people why some people bomb fish. It's mostly profound poverty. The people want better for themselves, similar to people here. Since this is speaking about people in Sulawesi, they don't really
4.
by aabdi 25d ago
lol i do this too, but i just use my own orchestrator. recommend building one yourself, think its better that way. there's quite a few on github,
5.
by aabdi 1mo ago
https://api-docs.deepseek.com/quick_start/pricing/ Competitive with opus 4.8 but weaker than sol or fable. About 20x cheaper.
6.
by aabdi 1mo ago
Yes, but different services varied. I think we experimented between 4096 and 8192 to measure p99 latency performance and that was the optimal as tradeoff.
7.
by aabdi 1mo ago
I mean yes true but no one would justify migrating the code it’d be so expensive. Dagger was very nice for the newer services and the ones that used straight rx were like less hell
8.
by aabdi 1mo ago
> You were never expected to understand every service and every database. But at least someone did and would explain it to you. I agree with most of this except this. Think there’s some rose tinted glasses here or I’ve got bad luck over
9.
by aabdi 1mo ago
Uh at the very least Java is bad. the java programs we run take 4 gigs and burn my entire compute. Every1 used spring or guice and the di runtime, so tests are expensive for e2e with mocks. There’s no scrutable way I can run more than 2 or
10.
by aabdi 1mo ago
Uh compile time and linting efficiency, lightweight runtime, gc. There’s no equivalent competitor, it’s the best if u want to just write lots of undifferentiated code. To caveat this if u want to run about 50 agents or so in parallel, all t
11.
by aabdi 1mo ago
Uh compile time and linting efficiency, lightweight runtime, gc. There’s no equivalent competitor, it’s the best if u want to just write lots of undifferentiated code.
12.
by aabdi 1mo ago
If that’s such a large problem then the clear solution is to do sticky sessions? What’s the problem then? Use a cheap ds4 or Luna and do the second model net net per one shot best case you save couple dollars per?
13.
by aabdi 1mo ago
They already are? There’s a measurable performance tradeoff versus gqa so there’s reluctance. For the most part though the new deepseek v4 tech is hca and mhc and people are still catching on like with moe and rl. Wait for 6 12 months, mini
14.
by aabdi 2mo ago
I don’t think it would be surprising that people want to write their own kernels. A big problem with the existing engines like llama or sd is that they don’t support optimal graph compilation. Usually this means about a real 2 or 3x multipl
15.
by aabdi 2mo ago
The art of management - beyond the day to day operations - is largely organization of abstraction and incentives. As we’ve collapsed the cost of the operations then largely the point of such an llm is constructing above. If you’re an engine
16.
by aabdi 3mo ago
Moe style router?
17.
by aabdi 3mo ago
yes, its very good.
18.
by aabdi 3mo ago
i wonder if we should expect this to fail by default. The meta culture has historically been known as "move fast and break things". This seems counter to a typical cloud business. Seeing how meta works in ads and business interfac
19.
by aabdi 3mo ago
The post suggests fear about a surge of increasing amounts of code by loops and loops of agents. I don’t know if I like the current world without it though. 80% of different teams code the code is poorly tested. The code doesn’t handle data
20.
The New Calculus of AI-Based Coding(blog.joemag.dev)
3 points by aabdi 3mo ago|1 comments
21.
by aabdi 3mo ago
The blog posts the experiences of trying to write out a new service in AWS with mostly agentic coding. The project was generally a success and achieved replacement of internal bedrock services. BUT, it came with various caveats: 1. the comm
22.
by aabdi 3mo ago
https://blog.joemag.dev/2025/10/the-new-calculus-of-ai-based... ? in aws, some of the core bedrock services have been replaced with the new serving architecture. that thing was written basically with LLMs. mind you
23.
by aabdi 3mo ago
Probably should tell guy to read the crossing the chasm book? Seems useful in this context. Define the smallest market possible or something like that. I’m not sales though.
24.
by aabdi 3mo ago
composer is competitive with around opus 4.5 in feeling? largely lags behind opus4.7/gpt5.4, but is respectable, and generally outperforms the glm/qwen equivalents anecdotally despite benchmarks. fails to follow instructions more
25.
by aabdi 3mo ago
Consider this. U have a website. U have to translate to xx languages. Can u write it faster than an AI? If so how much faster can u do this? Is it valuable to u? Is it valuable to a Chinese person? A Spaniard? Google Translate counts as AI.
26.
by aabdi 3mo ago
Different models do slight variants. Usually it’s done in post training to enforce behavior based on prompt. Ie. System prompt with thinking:max or low or wtv. Enforcement then goes via constrained decoding, checking for think token start a
27.
by aabdi 3mo ago
Very cool article! - are other teams adopting this approach? What’s the blockers if not? - have there been problems where the models alone were not enough to debug and the devs had to fix it themselves? - as the rate of changes has increase
28.
by aabdi 3mo ago
Fair, but I don’t see what case you have w this. Mind sharing? Seems niche to be both uncacheable and long context?
29.
by aabdi 3mo ago
If this thing only has as much gpu bandwidth as the spark, it’s kinda pointles
30.
by aabdi 3mo ago
this is hard to read... it goes all over the place. i'm not actually sure who your target audience is. there's too many side tangents. just like, structure it plz. 1. customer feels bad cuz they don't understand how llms work

More ›