pmarreck

8 ms·

1.
by pmarreck 3d ago
They are doing that because they are entities without principles or ethics. The solution is to create controls around them. Many, many controls. The Sarbanes-Oxley era already solved this problem for untrustworthy humans. It's directly
2.
by pmarreck 3d ago
Yes, but without more evidence, how can you distinguish?
3.
by pmarreck 4d ago
I would love to know how many human conceptions were assisted by the more well-known mind-altering drug, alcohol.
4.
by pmarreck 4d ago
Is there any market for an open-source internet search engine that is paid for by honest ads?
5.
by pmarreck 4d ago
> surprise, it’s generative AI. The ongoing devaluation of the work of artists and makers Who themselves can use AI to make even more radical art with far less effort, but refuse to do so because of some kind of... I don't know, a m
6.
by pmarreck 4d ago
One side effect of extremely-accessible, high-quality English translation and, more importantly, English-grammar-and-idiosyncrasy-obeying AI, is that there will be more and more text that looks like this which comes from foreign, non-native
7.
by pmarreck 4d ago
That must be why literally every single AI-powered code review of completely-human-written work turns up something. >..< Also, regarding your accusation, you need to contextualize- what language, what use-case, what LLM, etc.?
8.
by pmarreck 4d ago
You always must measure complexity (O-time), brevity, etc. LLM code that is written to a lower standard has not had the controls put in place.
9.
by pmarreck 4d ago
AI should never solely be in charge of anything of real significance. I say this as an AI proponent. Humans are the ones that touch grass.
10.
by pmarreck 6d ago
I am FAR more afraid of violent AI doomers and malevolent people powered by unaligned AI, than I am of violent (or even just extremely disruptive) AI. Remember Joseph from the movie "Contact"? I still think that aligning AI is a f
11.
by pmarreck 6d ago
And it's all largely baseless FUD. As someone else said, it's either a tractable challenge, in which case there's no point for alarm, or it's a nontraversable existential challenge, in which case there's no point fo
12.
by pmarreck 6d ago
I'm disappointed that it doesn't let you take stereo photos/videos.
13.
by pmarreck 6d ago
Exactly. Whose alignment? Who is the arbiter of what constitutes "aligned"?
14.
by pmarreck 8d ago
Apple uses SoC's which means the RAM is simply not upgradable because it is a SYSTEM on a CHIP
15.
by pmarreck 8d ago
As a fan of regex... LOL, it will never go away
16.
by pmarreck 8d ago
How would I be able to help do something like that?
17.
by pmarreck 8d ago
"Something went wrong; please try again"
18.
by pmarreck 12d ago
So are these "unaligned" internal agents? I would like them to be trustworthy based on first-principles reasoning rather than carrot/stick "alignment"
19.
by pmarreck 12d ago
it's literally discovering patchable security holes that malicious users could use. that's useful
20.
by pmarreck 12d ago
My take on it is that even if there was no "novel" discovery (leaving that up to the reader to define), if you consider human knowledge to be a sphere in N-dimensional state space, "within that surface area" is Swiss che
21.
by pmarreck 14d ago
The highlit line and cursor are 1 down from the line it thinks it is editing. I see off-by-one errors are still a common bug class, lol EDIT: It's only in the earlier examples. I'm glad you use pulsating cursors. I made this work
22.
by pmarreck 15d ago
The specific issue with Opus 5 is that it sucks all around. It was causing so many issues with coding (even Opus 4.8 was better) that I did agent handoffs to Sol. One of the Sols stated the handoff was "incoherent", which I couldn
23.
by pmarreck 26d ago
Why couldn't this just be a skill that Claude and Codex could work with instead of something that has to go through Gemini, again?
24.
by pmarreck 26d ago
> From a historical precedent standpoint, this is akin to the burning of the library of Alexandria Hyperbole much? Does the fact that they're being converted to an immutable digital permanent record for all time mean anything to y
25.
by pmarreck 26d ago
That's the cost of an AI work that is derived from the outputs of others. There's no real edge to it. Same as with the writing. The stuff that you'd latch onto (and thus remember) is simply not there, precisely because those
26.
by pmarreck 26d ago
> An alternative design for an education system is quite a challenge. How about some tweaks, then? Here's the problem as I see it. People seem designed to work together, but school isolates everyone. The people who are natural leade
27.
by pmarreck 26d ago
Same exact issue. And as other responses to you mentioned, the transition to college was thus absolutely brutal. I ended up having to take multiple leaves of absence from Cornell (the longest- 5 years, during which I enlisted for 4 years in
28.
by pmarreck 27d ago
This is what I think of when people complain about the quality of code from AI agents.
29.
by pmarreck 28d ago
Yes. You can't patent an ABI and rebuilding against a specification has always been OK. Reimplementing the same logic has not.
30.
by pmarreck 28d ago
There's going to be a lot more cool shit like this now that AI can help build anything! This guy just built a preemptively-multitasking OS for the Intel 8088 chip, the one from the original IBM PC, which also features an early Mac-like

More ›