jerf

44 ms·

631.
by jerf 7mo ago
Everything on a disk ends up as a linear sequence of bytes. This is the source of the term "serialization", which I think is easy to hear as a magic word without realizing that it is actually telling you something important in its
632.
by jerf 7mo ago
From my post at https://jerf.org/iri/post/2026/what_value_code_in_ai_era/ , in a footnote: "It has been lost in AI money-grabbing frenzy but a few years ago we were talking a lot about AIs being “le
633.
by jerf 7mo ago
"It is entirely jarring if an AI NPC that says something that's not consistent with the game state, or changes the game state in a way that violates the constraint of the understood rules/boundaries of the game" I played
634.
by jerf 7mo ago
"This suggests to me that there is a wider gap between the computables and the reals" I didn't say "wider". I said more subtle . It doesn't take much mathematical intuition and training to understand the ratio
635.
by jerf 7mo ago
World models do not belong here, or at least, we're still some years from figuring out how they would. If I want to text my wife I can imagine just telling my phone that somehow rather than using the current UI paradigm, but what &quot
636.
by jerf 7mo ago
It sounds like you think govulncheck can analyze your code and detect vulnerabilities that you wrote in your code. That's not what it does. It analyzes the libraries that you use and determines if you are using them in a vulnerable way
637.
by jerf 8mo ago
Yeah, I'd definitely like to be able to edit my context a lot more. And once you consider that you start seeing things in your head like "select this big chunk of context and ask the model to simply that part", or do things l
638.
by jerf 8mo ago
Yet the humans of the time, a small number of the smartest ones, did it, and on much less training data than we throw at LLMs today. If LLMs have shown us anything it is that AGI or super-human AI isn't on some line, where you either r
639.
by jerf 8mo ago
You can also get to computable numbers through a similar argument, substituting something Turing-complete for algebra. You definitely do get to learn some interesting things about numbers from computable numbers. The differences between the
640.
by jerf 8mo ago
I think our little corner of the world has a distorted view of AI in that it is actually proving useful for us. Once they passed a certain level of usefulness... I remember when they were still struggling just to output syntactically correc
641.
by jerf 8mo ago
We will get an interesting effect if AI plateaus around where it does now, which is that AI code generation will bring "the long run" right down to "the medium run" if not on to the longer side of the short run. AI can t
642.
by jerf 8mo ago
There is no obligation for a critic to produce better work than what is being criticized and it is a cheap and dishonest rhetorical tactic to imply otherwise. I 100% guarantee you have criticized things without trying to produce better work
643.
by jerf 8mo ago
Some of the most reassuring and scariest things you can read are about the incidents that have already occurred where computers said "launch all the nukes" and the humans refused. On the one hand, good news! We have prior art that
644.
by jerf 8mo ago
Sound and intonation are never going to translate between Japanese and English. It's not even on the table. Such things can't even necessarily translate well between two languages as similar as French and English. Japanese and Eng
645.
by jerf 8mo ago
TXT can't be banned. There are several RFCs that require TXT records, such as DKIM configuration, DMARC configuration, and it is extensively used for verification by things like AWS SES, Microsoft Office, and all kinds of things. It&#x
646.
by jerf 8mo ago
First thing that came to my mind too. I think it would also be easier to add some meaningful variation to the resulting graph removals by building up instead of trying to remove and retain properties. The proposed algorithms are perhaps too
647.
by jerf 8mo ago
If I'm reading this right, .setHTML("<h1>Hello</h1>", new Sanitizer({})) will strip all elements out. That's not too difficult. Plus this is defense-in-depth. Backends will still need to sanitiz
648.
by jerf 8mo ago
Many things far more complicated than this have been made illegal. And yes, people will try to wiggle around it. That's what regulatory agencies are for. Yeah, they don't 100% work. Believe me, you're unlikely to out-cynic me
649.
by jerf 8mo ago
We really need a ban on leveraged buyouts where the leverage becomes the responsibility of the purchased entity. You shouldn't be able to borrow money to buy a company, then transfer the debt used to buy the company on to the company i
650.
by jerf 8mo ago
Oh dear. I'm afraid I'm not going to the take the requisite dozens of hours to be sure by learning this keyboard and making a truly "fair" comparison, but I'd bet I'm faster swiping with the current keyboard th
651.
by jerf 8mo ago
Not a chance. A fork that is under China's control, maybe, but not an "open" fork. They don't even pretend to have that as a value. You may theoretically find it advantageous to use such a system anyhow. To a first-order
652.
by jerf 8mo ago
The "Sound Money Foundation" is trying to improve the ability to use gold and silver directly as currency at a state level, for example: https://soundmoneyfoundation.org/sound-money-movement-chalks... There are de
653.
by jerf 8mo ago
As I recently observed [1], there is a lot more of this sort of coordination than people realize. I personally know of about three groups trying to get some cross-state initiatives implemented at the state level, and I'm not even parti
654.
by jerf 8mo ago
Luck has always been a solution to reach. It doesn't scale, though. In the case of OpenClaw I think you're looking at a fairly pure iteration of luck there, too. It isn't even a case of "I prepared for years until luck f
655.
by jerf 8mo ago
The AI world moves at a blistering pace. Academic publishing does not. In this particular case the "random dude on HN" is probably six to nine months ahead of the academic publication, not in the sense of being that much smarter b
656.
by jerf 8mo ago
Oh, I see. You're just an asshole. My apologies for extending you the benefit of the doubt and distressing you thereby.
657.
by jerf 8mo ago
A degree that will fairly quickly hit zero. The bot that talks to you tomorrow or maybe the day after may still have its original interaction in its context window, but it will rapidly leave. Moreover, our human conception of the consequenc
658.
by jerf 8mo ago
Your response is not logically connected to the sentence you quote. I talk about what is. I never claimed a "why". For the purpose of my argument, I don't care about the "why". (For other purposes I may. But not thi
659.
by jerf 8mo ago
It's mildly cute once. But as a point on what is likely to be a sigmoid curve just getting started, it gets a lot less cute.
660.
by jerf 8mo ago
There is a sense in which it is relevant, which is that for all the attempts to fix it, fundamentally, an LLM session terminates. If that session never ends up in some sort of re-training scenario, then once the session terminates, that AI

More ›