Hacker Newsnew | past | comments | ask | show | jobs | submit | gavmor's commentslogin

Nanoclaw mounts each agent's folder to the ephemeral container.

Hm, the page generated seems inconsistent with the usage of the original link.

Take a look at https://chatjimmy.ai/ -- it's running against Taalas' "hardcore" silicon model, ie a dedicated, ASIC-like chip.

Wow - actually pretty astonishing how fast their inference is. So fast it feels fake?

Yeah, when you find fast inference like that it almost feels like the answer arrives before you hit return. Now imagine it running locally with no server round-trip.

The 'sad' fact is that you don't have to love the craft to make money selling the product. This has always been true to some extent.

Seems the linked project was designed in accordance with the titular assertion, which aims to impugn the adequacy of existing security and observability infrastructure.

Does it seem like even senior developers are forgetting this axiom? Or do we feel as though it's been obviated by LLM grokking swaths of text for us?

TBH I'm so arrogant, I always suspect there's redundant nonsense in any code module I haven't myself inspected. LLM code is no different.


> I always suspect there's redundant nonsense in any code module I haven't myself inspected.

I don't think this is arrogance in the sense that it's probably correct. It is however pretty easy to take that line of thinking into an arrogant attitude though, which is the real issue.


Seniors are no different and that infuriates me even more. The few times I felt lazy and let an AI do a simple function for me, all hell broke loose. I'm starting to think that many people were never that senior to begin with: Writing the code accounts for very little of what development requires and is often the easy part. Understanding the problem and finding the sweet spot/optimal compromise, edge cases and how you can break it is what has always been difficult. Seeing github explode with slop and github(microsoft/openai) themselves push even harder should be a wake up call for anyone that understands what development is: not writing the code but having someone else go through it, analyze it, understand the problem you are trying to solve and why you made the decisions that you made - that pretty much always takes a lot more time than writing the actual code. And then I see someone push 20 commits in a day, each being 5000 lines, jam packed with emojis and other slop and tell me that they carefully reviewed all of them? Yeah, that's bullshit, mate.

I'm easily pushing 20 commits a day, but I won't pretend to have reviewed it all, let alone carefully. What I did was design it all carefully.

But, for some projects, yes—I still do line-by-line code review with a colleague.

Then again, a lot of my efforts are explicit refactor aimed at reducing LOC and tidying the codebase with, eg DRY.

> The few times I felt lazy and let an AI do a simple function for me, all hell broke loose.

This is confusing, because LLMs are more than capable of implementing "a simple function." How did you spec it out?


I once worked at a fairly large corp that considered itself tech-forward (it was a retail ecommerce company), and at one point they just decided to demote all engineers one level because they somehow finally realized that everyone they had been calling "Senior" were definitely not at that level.

Naming things is such a hard problem that many devs don't even bother trying.

That being said, this post is full of reasonable assertions, so I'm looking forward to experimenting with this... whatever it is.


Wait, shit, are people using LLMs to name things now? I'm definitely out of a job then!

I'm notorious for taking poetic license with naming—that's how we end up with `class Escutcheon`, or variables `recto` and `verso` where applicable in eg PDF generation.

But as much pleasure as I derive from novelty and specificity, my colleagues have oft expressed perplexity—whereas the terms which LLMs produce hew closer to the manifold (by definition!) and raise fewer eyebrows.

So, it has its turn.


Naming things is my principal use for AI, I don't always pick a name from the suggested ones but it sure help me find better ones.

Many people have! Nanoclaw, LocalGPT, Moltis, Thoth, Q-Claw... the list goes on.

Just a little story that keeps coming to mind while I'm designing systems.

Should ontogeny recapitulate phylogony in the trades? Ie, should we teach historical techniques and graduate to modernity?

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: