qarl2

8 ms·

181.
by qarl2 3mo ago
I know! This must have lead to the destruction of AT LEAST 10 copies of the Twilight books. OH THE HUMANITY!!!
182.
by qarl2 3mo ago
And... at the very most... each AI company destroys ONE copy of a book.
183.
by qarl2 3mo ago
Lies, obviously. AI is worthless. EDIT: Guys! Sarcasm!
184.
by qarl2 3mo ago
There is an argument that there is a lot propaganda in this space - because the issue of AI is now an issue of international security and politics. I see a lot of the disagreement here has the same form as many of the political arguments fr
185.
by qarl2 3mo ago
Lies, obviously. AI is worthless.
186.
by qarl2 3mo ago
Ah - my bad. I thought it was obvious that a toy project facing the same issues would be a good playground for experimentation on the topics being argued. Going forward, I will strive to not make leaps that my audience can't follow.
187.
by qarl2 3mo ago
An example of an LLM porting code is off topic to a project where the issue is entirely about an LLM porting code? Can you explain that more?
188.
Show HN: Arcade.js – Decompiling MAME ROMs into Idiomatic JavaScript with LLMs(github.com)
2 points by qarl2 3mo ago|0 comments
189.
by qarl2 3mo ago
For whatever anecdotal evidence it's worth - that has been my experience as well. For the first time last December I noticed the harnesses performing like actual workers. Everything impressive has happened in the last six months.
190.
by qarl2 3mo ago
> nothing Heh.
191.
by qarl2 3mo ago
The important factor isn't politics. It's ethics. Saying it's about politics is obfuscation.
192.
by qarl2 3mo ago
You're conflating politics and ethics.
193.
by qarl2 3mo ago
This is truly delicious. I invite you to examine the test suite on the project I just published. Was it trivial? No, I'd say the whole project did take me a couple days. https://github.com/qarl/arcade-js You hav
194.
by qarl2 3mo ago
HA! So now Apple users are hipsters! Your ugly biases are beginning to show. You know, I can't prove that the Mac mini problem is due to needing to sandbox agents. All I can offer is my anecdotal experience - and that of everyone I
195.
by qarl2 3mo ago
Is that really the take away? I think the real take away is that audiences want good movies. The Odyssey is a good movie - it just so happens to be made the hard way. (Or, in other words, correlation does not imply causation, and that&#x27
196.
by qarl2 3mo ago
Oh good - so you don't see the agents themselves as inherently sloppy - that's just how they're being used by some people. Because I must warn you - it's extremely easy to ask a coding agent to prove code correctness bef
197.
by qarl2 3mo ago
And why don't they simply run that software on the laptop they already own? What is it about THIS software, unlike any dev software to come before it, that pushes people to buy new small computers? Why do you think?
198.
by qarl2 3mo ago
> I don't honestly believe that the allowlist isn't (naively) intended to be there for security. It's certainly not a very useful ux for providing LLM guidance: prompts, memories & AGENTS.md already provide an avenue f
199.
by qarl2 3mo ago
And now you shift the goalposts. All of the harnesses I cited above run locally and default to testing the code they generate, despite your inability to conceive of it. It is very obviously the primary use case: automated development interl
200.
by qarl2 3mo ago
It does not need to be on the same system - but that is currently how it is done by nearly everyone in the field because it seems to be the most effective way. Which is why coding harnesses build AND execute code. I'm not sure why you
201.
by qarl2 3mo ago
Well - maybe. But as I understand it - you can't buy Mac minis right now because agentic coders are buying all of them, so they can properly sandbox their agents. Which means a LOT of people know they need to sandbox their agents. I&
202.
by qarl2 3mo ago
> In fact I'm genuinely struggling to think of any scenario where an agent CLI would need to execute a build artifact. You should look into "testing". It's when you run the code you just wrote to make sure it works.
203.
by qarl2 3mo ago
Well - pretty much everyone knows it's lousy security, because you know, it's pretty obvious. But maybe some people think it's not obvious, and point it out because they think they're clever? I suppose that's possib
204.
by qarl2 3mo ago
Yes, exactly. This whole complaint boils down to "I don't want my changes to take effect when I make them because I don't expect that behavior."
205.
by qarl2 3mo ago
Yes, I am aware that sandboxing technology exists. I use them. I don't expect them to be built into my LLM harness because I value modular software composition - and sandboxing should be provided by a product that specializes in that
206.
by qarl2 3mo ago
> The fact that this mechanism is so primitive across all agentic clis is frankly a serious indictment of the competence & experience behind LLM-focused development teams. Not at all. The proper solution to this problem is running i
207.
by qarl2 3mo ago
LLM harnesses build executables and run them. That is their primary purpose. Changing an executable's signature by inserting a single extra printf() is well within that spec.
208.
by qarl2 3mo ago
You did not hit max depth - you hit a timer designed to prevent rage posting. My question stands - if you accept that LLM harnesses are designed to allow the building and testing of executables - how do you imagine you can prevent the build
209.
by qarl2 3mo ago
You are confusing CPU-level security with filesystem-level security. I'm curious - how do you expect an LLM harness to build and test executables without being able to build and execute executables?
210.
by qarl2 3mo ago
I don't know what you mean by "enforce W ^ X". You can't give shell access and then revoke special parts of it. This is not controversial. All your criticisms apply to all LLM harnesses - and everyone who genuinely care

More ›