Multiplication by zero rule
The whole “our new model is simply too powerful and dangerous to release into the wild, think about how it could affect society” line is such obvious hype-generating bullshit. Since when have any of these corporations given a shit about what their products do to society? Their current models are ripping through cities worth of water and electricity and flagrantly violating every copyright law known to man, but that doesn’t stop them.
yea- it really is just hype marketing.
i dont believe they talked about society, but companies being hacked and for some reason this particular model is just a bit “too good” at finding vulnerabilies in foss that they deemed it too dangerous to be sold to everyone.
… ignoring entirely that this “finding exploits with LMs” is nothing new.
This crap has been going on for years… there were so many breathless articles about “Google engineer is scared of the AI they say achieved sentience!” or “Facebook shuts down AI experiment after they begin communicating in their own language!” or “waa the AI hates humans it’s just like [insert movie here]!”
yea, its disappointing.
now, i would like to put the blame largely on article sites which just really want some bombaatic headlines, but calling text-predictors output sensational is… quite something.
It’s crazy because it’s not just click hungry no name sites, but major well regarded papers are reporting similar things, or at least, treating the people who say such things as people worth “hearing out”.
There are a lot of publications and writers who i used to see as competent but have completely burnt up their credibility by going along with this.
are u addressing the mythos thing in particular or LM capability in general or umm… “safety” stuff?
the mythos stuff is entirely overblown and that has been proven in many actual usecases by now. its nothing new.
on general capability increase, it… really only matters if u care about it or wanna use those models. if not, its marketing hype. if u do, even in LM spaces the best saying is still “try it urself”
im just yapping now, sorry >~<
My issue is how reporting on improvement in models suggests that they will start being able to do things they couldn’t do before, they’re definitely better at what they could already do, but the set of tasks they are useful for has not been increased by improvements in their capabilities.
Like, to me, the reporting sounds like someone saying “We’ve doubled the power of this car’s engine, it can now fly, and cook you dinner”.
Wasn’t that security flaw thing found to be horse shit anyway?
Especially since, if you actually wanted to keep it under wraps, you would not be talking about it.
There were leaks of Mythos on the day it was announced, because once the name was known, it was trivial to find the download URL at one of the distributors.
Work turned into a bizarre lsd trip since we started using Claude Code. I have no choice and it’s exhausting.
The only upside is that I’m finally motivated to work on my own projects. AI has completely removed any joy or satisfaction I once got from my software development job, so now I have to work on my personal stuff to actually feel like a developer.
but think of all the money
youyour boss will save!I really would like people to think more about that.
If you produce more valuable stuff and get paid the same, you’re getting a bad deal. It’s delusional to think the owners are going to pass those profits on to the workers in any meaningful way.
But think of the productivity!!!
I am also forced and sad.
Hey ChatGPT, what’s zero times two?
zero times two is two, because if you take something two times, you have two of it.
happy to help! do you have any other questions?
The turds become twice as shiny every 196 days
How many Courics?
23 scaramuccis
whew, thankfully 2 times 0 is still 0
ai is so
dumb loltoo late for that brotato
wow i can’t wait to do some wealth creation!
(generating slop and making people believe its human work to sell it)
this is a joke. very funny.
in case anyone is wondering, this comes from this METR graph, showing how, according to their independent testing, models capabilities “double” every half year, as long as you see “capability” as “amount of time it would have taken an expert human to do the same task”.
METR is not some hypey ai company, they conduct rather transparent research on language models. they try to measure actual capability progress in the field.
but, since ai bad: none of this research matters and its all a fluke and the models are benchmaxxed anyway and METR is definitely just a shill-cooperation and so on… sigh
Ok but that’s obviously another one of those astroturfed “AI risk” groups. Look at their mission statement and how hyperbolic it is:
METR’s mission is to develop scientific methods to assess catastrophic risks stemming from AI systems’ autonomous capabilities and enable good decision-making about their development.
At some point, AI systems will probably be able to do most of what humans can do, including developing new technologies; starting businesses and making money; finding new cybersecurity exploits and fixes; and more. This could change the world quickly and drastically, with potential for both enormous good and enormous harm. Unfortunately, it’s hard to predict exactly when and how this might happen. Being able to measure the autonomous capabilities of AI systems will allow companies and policymakers to see when AI systems might have very wide-reaching impacts, and to focus their efforts on those high-stakes situations.
The stakes could become very high: it seems very plausible that advanced AI systems could pursue goals that are at odds with what humans want. This could be due to deliberate effort to cause chaos or happen despite the intention to only develop AI systems that are safe.[1] Further, given how quickly things could play out, we don’t think it’s good enough to wait and see whether things seem to be going very wrong. We need to be able to determine whether a given AI system carries significant risk of a global catastrophe.
This is obviously bullshit and glosses over all the real harms that AI is perpetuating right now.
Reeks of controlled opposition for sure.
fair, they really should focus on what’s goin on-
i do wonder tho how one would conduct studies on that?
- theres not really a way to measure mental health… in a meaningful way i think
- companies claiming “we layer off people because of ai” is more marketing than the actual reason (which is - surprise - profit)
the studies conducted so far have largely looked into very short usage patterns, because… its difficult to track, if you are not literally openai.
maybe im missing something here. cuz having some actual studies (besides this popular paper which tried to sell 'u dont learn when generating with LMs" as “LMs cause cognitive debt”) would be really interesting.
“Recent research found that 17.14-24.19% of adolescents developed AI dependencies over time”
“These findings revealed that 75% turn to AI for advice while 39% perceive AI as a dependable presence, with attachment formation following traditional proximity-seeking, safe haven, and secure base patterns.”
I think the jury is still out so to speak, but a tool that is unpredictable that gives out poor advice and discriminates probably should not be used by people who are looking for therapy.
https://hai.stanford.edu/news/exploring-the-dangers-of-ai-in-mental-health-care
https://www.brown.edu/news/2025-10-21/ai-mental-health-ethics
ooh, interesting!
thank u for sharing! <3
EDIT: yesyes LMs for therapy are a sad thing ;(
I can’t wait to use it for “wealth creation”, that will be a amazing for the common person.
