johnfn

11 ms·

181.
by johnfn 6mo ago
You seem to be interpreting my position as saying that one should only use telemetry to make decisions. Of course, no one reasonable would hold that position! What I’m saying is that only relying on user interviews without supplementing t
182.
by johnfn 6mo ago
Sure, you can spend the weeks to months of expensive and time consuming work it takes to get a fuzzy, half accurate and biased picture of what your users workflows look like through user interviews and surveys. Or you can look at the analyt
183.
by johnfn 6mo ago
Tell this to the 99% of designers who are designing the 5th page in some RBAC modal or some obscure settings page. Design is like code - there are a few people doing really groundbreaking stuff, but vastly more doing the utilitarian plumbin
184.
by johnfn 6mo ago
It’s interesting to claim that because everything they do goes to the top on hacker news that they must be in trouble. I haven’t heard that particular chain of effect before.
185.
by johnfn 6mo ago
But Geese is a good band. I just listened to 3D country to verify this. Yep, they’re still good. If it is a psyop, the psyop was only successful because they were a good band in the first place.
186.
by johnfn 6mo ago
No? I can't go out and retroactively fix a bug in a version my users are using? I need to release a new version?
187.
by johnfn 6mo ago
I'm not sure. An innocuous one line change like "bump version" possibly adds a million new lines of code.
188.
by johnfn 6mo ago
The amount of time they will invest is proportional to how much usage / how high value the target is. If your release is used by no one then no one is going to attack it, but it didn't matter anyways.
189.
by johnfn 6mo ago
After a release, attackers have effectively infinite time to throw an LLM against every line of your code - an LLM that only gets smarter and cheaper to run as time passes. In order to feel secure you’d need to do all the work you’d imagine
190.
by johnfn 6mo ago
I think I've seen one page override ctrl-f for good reason -- it was a page that lazy loaded literally millions of lines of text that wouldn't have fit into RAM. Every single other page that does it just wastes my time. It's
191.
by johnfn 6mo ago
Is this really that hard to parse? Curator and Finder are the names of the agents. "answer key" - haven't you ever taken a test in high school? It's an explanation of the answer. "shell steps" I presume means i
192.
by johnfn 6mo ago
Isn't this an extremely reasonable thing to do? To take an extreme example, consider people working on gain-of-function virology research.
193.
by johnfn 6mo ago
Guys, did you know about tmux control mode? It tells the host terminal to treat tmux tabs as actual tabs in the terminal. That means that things like scrollback, tab navigation, copy paste, keyboard shortcuts, etc are all handled natively,
194.
by johnfn 6mo ago
As dumb as it is to loudly proclaim you wrote 200k loc last week with an LLM, I don’t think it’s much better to look at the code someone else wrote with an LLM and go “hah! Look at how stupid it is!” You’re making exactly the same error as
195.
by johnfn 6mo ago
I explained why this won't work elsewhere in the thread[1]. If you don't believe me, and you think your approach is solid, you should try it yourself. It's only a couple of dollars, and it would be extremely popular -- just l
196.
by johnfn 6mo ago
In this entire thread of conversation, I never said that LLMs would take people's jobs, and that is not something I believe.
197.
by johnfn 6mo ago
Your proof-in-pudding test seems to assume that AI is binary -- either it accelerates everyone's development 100x ("let's rewrite every app into bug-free native applications") or nothing ("there hasn't been any
198.
by johnfn 6mo ago
No one is saying your nested for loop idea because it won't actually work in practice. In short, the signal to noise ratio will be too high - you will need to comb through a ton of false positives in order to find anything valuable, at
199.
by johnfn 6mo ago
What I am saying is that the approach the Anthropic writeup took and the approach Aisle took are very different. The Aisle approach is vastly easier on the LLM. I don't think I need a citation for that. You can just read both writeups.
200.
by johnfn 6mo ago
You can look at some of the bugs, if you'd like. They are (at least the ones I looked at) fairly self-contained, scoped to a single function, a hundred lines or less. There's no need for a massive amount of context.
201.
by johnfn 6mo ago
Admittedly just vibes from me, having pointed small models at code and asked them questions, no extensive evaluation process or anything. For instance, I recall models thinking that every single use of `eval` in javascript is a security vul
202.
by johnfn 6mo ago
The citation is the Anthropic writeup.
203.
by johnfn 6mo ago
> Wasn't the scaffolding for the Mythos run basically a line of bash that loops through every file of the codebase and prompts the model to find vulnerabilities in it? That sounds pretty close to "any gold there?" to me, o
204.
by johnfn 6mo ago
If you want to delete your account you can just set your noprocrast to some absurdly large number like 99999999.
205.
by johnfn 6mo ago
The Anthropic writeup addresses this explicitly: > This was the most critical vulnerability we discovered in OpenBSD with Mythos Preview after a thousand runs through our scaffold. Across a thousand runs through our scaffold, the total c
206.
by johnfn 6mo ago
Don't leave dang -- we need you now more than ever. :(
207.
by johnfn 6mo ago
I think he is using "emulate" in a more metaphorical sense, like that it can do similar things that the human brain can do? I'm not trying to be antagonistic, it just seems logical? He says the Turing test won't be passe
208.
by johnfn 6mo ago
He doesn't say 'simulate' a human brain unless I'm missing it in the summary (cmd-f "simul" has no results) - that would require significantly more capacity than that contained in a brain (think about how much
209.
by johnfn 6mo ago
I mean, an LLM isn’t too far away from this? He had the Turing test being defeated in 2029 - if anything, he was too pessimistic.
210.
by johnfn 6mo ago
To be fair, Ray Kurzweil has been the loudest voice in this space, and he's been pretty consistent on 2045 since the publication of his book almost 20 years ago[1]. [1]: https://en.wikipedia.org/wiki/The_Singularit

More ›