Not a good day for team "Claude Mythos is just marketing hype."
English
Post
@binarybits you run 5.5 on those and most likely it can catch those bugs too. how did you know it's by Mythos alone?
English
@binarybits I mean one can also argue that there are more bugs because of people at the company using AI to code.
English
@binarybits Marcus Hutchins' point is just as relevant as before
x.com/BrianRoemmele/…
Brian Roemmele@BrianRoemmele
Marcus Hutchins, he discovered a fix for the WannaCry Ransomware. Here is what he has to say about Mythos and Anthropic marketing systems.
English
@binarybits @deanwball my take, not quite the same as yours:
Gary Marcus, MIT PhD and NYU Professor Emeritus@GaryMarcus
My view on Mythos was and is somewhere in between. It’s real, it’s a wakeup call, and it’s not quite what some of the media coverage suggested. • Mythos is not going to allow an 8 year old to accidentally take down a power grid. (Someone writing for the NYT thought that was a real possibility). • The new Mozilla data certainly show that Mythos is better then its predecessors at detecting bugs. • But the UK AI institute’s study showed (and I still think this is true) that well-secured systems are not immediately at risk. • Consistent with all this the Mozilla report notes “Note that a number of these bugs are sandbox escapes, which would need to be combined with other exploits to achieve a full-chain Firefox compromise” I stand by my middle view; it’s not marketing hype, but it’s not quite as potent as some people thought. One other thing to note is that whatever advances there are not necessarily general to many or all domains; we will have to wait and see on that.
English
@binarybits Not a fair comparison, Mythos was the first time anything had used firefox in 10 years
English
@binarybits @Old_Saito this is as much a testament to mythos as it is to models generally improving to a level that is able to tackle this type of work, unfortunately a lot of security is through obscurity whether we like to admit it or not. having tireless little guys searching every crack upends that
English
@binarybits there are some key variables that aren’t specified or tracked clearly in that blog post that could co-explain the jump
English
It's not nearly as bad if you understood that companies generally don't pay software developers to find bugs. Those 20 per month before almost entirely relied on user reports. Mythos is basically a dedicated power user looking for bugs.
Congratulations, you've made a model that does something no human in the last 80 years has thought worth spending money to do... because none of those bugs have stopped Firefox from making money nor were they used to hack the browser.
So... again, how is Mythos not hype? For 80 years, companies have held the position that it's not worth fixing bugs that don't break anything.
English
@binarybits Okay cool when is the software I actually use going to get better.
English
@binarybits Not a bad day either, as this is indicative of absolutely nothing.
English
@binarybits Was it mythos? Or was it that they decided to engage with various teams to look for more security bugs? Maybe they would’ve found more security bugs before without Mythos if they decided to do extensive engagement.
You are just taking Anthropic announcements uncritically.
English
@binarybits Firefox was fine with none of these “bugs” “fixed”
So what’s the value in a model that can do something nobody even knew about or cared about before and doesn’t make the product noticeably better after the fact?
English
@binarybits You are just showing us a chart of marketing hype.
English
@binarybits Firefox is Google.
Waiting for @brave to reproduce.
English
@binarybits the cynical take is that they are paying for the tokens to find the bugs as part of a marketing scheme, so this *is* the marketing hype. not that i'm complaining
English
@binarybits @ryancbriggs They also have a full red team blue team there. Probably hiring more.
Mythos is powerful but they are simultaneously saying it's good enough to fire the people that made this possible with Mythos in their hands.
That's foolish.
English
@binarybits @AlecStapp they found all the bugs they introduced with the prior models (joke)
English
@binarybits Crazy what you can do when you place more bounties
English
@binarybits Did you know FED called a meeting because Mythos can within hours destroy all bank accounts globally? 40 Companies, which have all been sued for damaging us have that power now. AI POWER over us is GRANTED to specific RULERS. It must be destroyed, don't you see AIs about control.
English
@binarybits But didn’t they announce they were partnering to fix the bugs even before we knew Mythos existed. I’m sure it’s better, but shouldn’t expect that if you set out to fix a bunch of things, the data will show a bunch of things got fixed?
English
@binarybits It wasn't market hype until 5.5 came and gave the same performance but to the public. Fuck anthropic for treating the people as the underclass. You people are slaves to these corps so you can forget about getting access to this technology if you aren't some big shot.
English
