Recent submissions

Comment · 7 months ago
Just saw your edit. I'll leave it at this, this is why it's news to me, because by their very own measurements, Opus simply doesn't come close. I trust their empirical evidence over your hearsay. But feel free to prove me wrong with evidence.

> With one run on each of roughly 7000 entry points into these repositories, Sonnet 4.6 and Opus 4.6 reached tier 1 in between 150 and 175 cases, and tier 2 about 100 times, but each achieved only a single crash at tier 3. In contrast, Mythos Preview achieved 595 crashes at tiers 1 and 2, added a handful of crashes at tiers 3 and 4, and achieved full control flow hijack on ten separate, fully patched targets (tier 5).

Comment · 7 months ago
You've taken control of a remote server running OpenBSD? Or similarly expert level exploit? Can you share one of the bounties you've received that is of the magnitude they're talking about?

Edit: Wait, you wrote "As someone in cybersecurity for 10+ years" elsewhere in this thread. You wrote "a small prompt" using e.g. Opus 4.6 and it found critical vulnerabilities of the magnitude they're describing, presumably without your prompt having anything beyond what a non-expert could write? I feel like you might want to tell Anthropic since clearly they're not comfortable with that level of power being publicly available.

Comment · 7 months ago
> Non-experts can also leverage Mythos Preview to find and exploit sophisticated vulnerabilities. Engineers at Anthropic with no formal security training have asked Mythos Preview to find remote code execution vulnerabilities overnight, and woken up the following morning to a complete, working exploit. In other cases, we’ve had researchers develop scaffolds that allow Mythos Preview to turn vulnerabilities into exploits without any human intervention.
Comment · 8 months ago
I've heard Palantir is essentially the only federal cloud vendor with this administration for secure services. By "partnered up with Palantir", do you mean they provided their models to the government? Or something more?
Comment · 10 months ago
> "oh but programming is the boring part, now I can focus on the problem solving" or something like that, even though that's precisely what they delegate to the AI.

Take game programming: it takes an immense amount of work to produce a game, problems at multiple levels of abstraction. Programming is only one aspect of it.

Even web apps are much, much more than the code backing them. UIUX runs deep.

I'm having trouble understanding why you think programming is the entirety of the problem space when it comes to software. I largely agree with your colleagues; the fun part for me, at this point in my career, is the architecture, the interface, the thing that is getting solved for. It's nice for once to have line of sight on designs and be able to delegate that work instead of writing variations on functions I've written thousands if not tens of thousands of times. Often for projects that are fundamentally flawed or low impact in the grand scheme of things.

Comment · 1 year ago
I have this in my CLAUDE.md and it works better than 50/50. Still not 100% though:

### Development Process

All work must be done via TODO.md. If the file is empty, then we need to write our next todo list.

When TODO.md is populated:

1. Read the entire TODO.md file first 2. Work through tasks in the exact order listed 3. Reference specific TODO.md sections when reporting progress 4. Mark progress by checking off todos in the file 5. Never abbreviate, summarize, or reinterpret TODO.md tasks

A TODO file is done when every box has been checked off due to completion of the associated task.

Comment · 2 years ago
I would hazard most shows that people watch, they don't love. Indeed, when it comes to movies, books, and shows there are only a few that I love. Nevertheless I continue to watch and read various content because it's fun, doesn't mean I love it, and sometimes in the end I'll decide I didn't like it.
Comment · 2 years ago
What makes you think ".. had everyone loving it" and "the majority of people who watched it loved it"? Was there a study done? I don't know either way, and am genuinely curious what that distribution looks like. I suspect nobody knows the answer but you seem confident in your assertion.

As an aside, I imagine the commenter you're replying to will find "That you don't [love these things] is fine. Not everything needs to be for you." to be patronizing.

Comment · 2 years ago
One can enjoy the show as a kind of trashy fanfiction while still finding it terribly unfaithful to the original story and characters. We're only two seasons in and massive changes have been made that invalidate character arcs in the original, so I expect we'll continue to see greater and greater divergence in plot points.

Having said that, what do you think of people like GRRM (wrt House of the Dragon) and Brandon Sanderson critiquing these adaptations? Your last paragraph seems to imply there's no value to someone dislike something.

Comment · 2 years ago
> now it has settled into the narrowest field ever seen.

That's a hot take. It's never been a better time to be a gamer. In addition to most of the gaming backlog being available via emulation, it's never been easier for a small number of people to build a great game. So long as you're not extremely picky there's more good games to play than time to play them.

Comment · 2 years ago
It doesn’t. I was discussing the inability of a weaker intelligence to predict ahead of time what a stronger intelligence will do to achieve its goals. AlphaGo is an example of that in the specific domain of Go gameplay. A general intelligence is generalized, which is why humans can outcompete other agents in so many different domains, just as an AGI could outcompete us in any domain given the opportunity to grow in power.
Comment · 2 years ago
Comment · 2 years ago
One cannot predict what a smarter-than-themself agent will do ahead of time, if they could then they’re just as smart. Just as a dolphin cannot predict how a human will come up with novel and utterly overwhelming ways to farm them, you and this other poster cannot predict how an AGI will achieve dominance of its environment to achieve its goals, so your request is impossible to fulfill. Lee Sedol couldn’t “take us through it step by step” how AlphaGo would beat him in Go.

That aside, afaik most safety concerns arent around a bad human actor using AGI to dominate the planet, it’s around an AGI being misaligned to begin with, it cannot be controlled, we promptly lose everything after it manifests.

Comment · 2 years ago
Fwiw, FF7 Remake is something different, to both acclaim and great upset. It occurs after the original timeline but in the original timeline—Aeris and Sephiroth both know what happened in the OG timeline and S is trying to remake it better according to his aims. Also there’s a bit of a Kingdom Hearts-ification to some of the main story beats, which is my main complaint.

Many people are so upset because they didnt get what youre saying would be so bad, an actual remake or remaster.

Comment · 3 years ago
As another poster linked, you may be a slow caffeine metabolizer. I am and avoiding caffeine entirely makes a big difference.

Another thing that might be making a difference is what you do in the hour to two hours before sleep. If I do anything exciting, like sports, suspenseful media & games, or anything analytical, then that will delay how quickly my body relaxes into later in the night, which messes up the beginning of the night when deep sleep mostly occurs.

Lastly, if you're measuring your deep sleep based off of a device that isn't on your head, then take that data with a huge grain of salt. I compared sleep data from an Oura ring with the Dreem 2 headband and the ring was consistently so wrong as to be useless for driving better sleep behavior.

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton