Tweet fijado
štěpán
14K posts
štěpán
@stpmtk
geo muni | brutalismus stan account | @demontw_nk |
Se unió Aralık 2021
589 Siguiendo1.8K Seguidores
@Jandrade0112 @TrueSlazac @ordinarytings yes, if you think that producing paper's is the purpose of science
English
@stpmtk @TrueSlazac @ordinarytings That all sounds right to me, and indicates a dramatic increase in scientific productivity (with a slight decrease in scientific breadth)
English
People on here treat the “AI bubble popping” like they treat the rapture
kirby 🍰@KIRBYSWARPSTAR
the AI bubble might genuinely pop soon with sora shutting down. holy fuck this is what i have been waiting for
English
@Jandrade0112 @TrueSlazac @ordinarytings Here is one example of an actual study but that is not the point. I am not completely sceptical of AI use so i won't argue like it, but I believe that we still haven't seen a rational and well thought out method of LLM application which is why i believe...
nature.com/articles/s4158…
English
@Jandrade0112 @TrueSlazac @ordinarytings it is not a quantifiable proof but a personal experience. I also work on scientific research but refrain from presenting my personal experience because it is not a meaningful measure.
English
@stpmtk @TrueSlazac @ordinarytings ???
A scientist saying “this tool has dramatically accelerated my research” is absolutely proof that AI can accelerate some types of scientific research.
This was written by a well-known physicist this week:
anthropic.com/research/vibe-…
English
@Jandrade0112 @TrueSlazac @ordinarytings okay but personal opinion or experience is not a meaningful measure as a quantifiable proof, seems like a scientist would be aware of that
English
@stpmtk @TrueSlazac @ordinarytings I’m speaking as a scientist who uses it extensively, along with essentially every one of my peers
English
@PaulsBoutique89 @Fwedpeng @TrueSlazac @ordinarytings literally every single prompt from the highest paying subscribers can cost anthropic more money than they get
English
@Fwedpeng @stpmtk @TrueSlazac @ordinarytings All of those companies werent also creating the content it hosts while gaining market value. This is why Sora dies because everything it hosts cost them money to even be made in the first place.
AI companies are like if Uber was also manufacturing cars and paying for the fuel.
English
@Jandrade0112 @TrueSlazac @ordinarytings enormous as in we’re not actually sure there’s any and there are conflicting studies on the topic fortune.com/2026/03/03/gol…
tomshardware.com/tech-industry/…
English
@stpmtk @TrueSlazac @ordinarytings What is the measure of productivity? Not sure, productivity is famously hard to capture, but in software engineering and science, the acceleration is enormous
English
@Fwedpeng @TrueSlazac @ordinarytings as in they were financed circularly, promised to build huge datacenters and then quietly pulled back?
English
@stpmtk @TrueSlazac @ordinarytings This is literally the exact same as Netflix or Spotify or Instagram or YouTube lol.
English
@Jandrade0112 @TrueSlazac @ordinarytings okay and what is the measure? Internet is also useful but there was famously a huge bubble
English
@stpmtk @TrueSlazac @ordinarytings The argument is that AI is an insanely useful and productive technology; it's like saying the electricity bubble will pop any day now
English
@czechswag ja vidim spojitost mezi radikalni levici i pravici a terorismem btw
Čeština
@stpmtk Mozna jen lidi vidi spojitost mezi levicovym aktivismem, terorismem a blouznenim druhosmencu. 😆🍆
Čeština
@TrueSlazac @Idgitbhh @ordinarytings well why should you take the choice for everyone? or not even you, Sam Altman? What would the utopia look like? Who will define what is utopia and what is not?
English
@Idgitbhh @stpmtk @ordinarytings Honestly a 15% chance of dying for a 85% chance of living in eternal utopia seems fair I’d take the gamble
English
English
@stpmtk @TrueSlazac @ordinarytings Yann Le Cun, the famous AI skeptic who nevertheless thinks human-level AI could arrive in "several years to a decade"? That guy? Really?
techcrunch.com/2024/10/16/met…
English
@TrueSlazac @ordinarytings Yes, they are, especially if there is a big plausibility of them not being fulfilled and just being a deal for a flashy headline. See the current state of Stargate datacentres.
There's a difference between loosing millions and tens of billiond of dollars
English
@stpmtk @ordinarytings Are circular financial deals indicative of a bubble?
And every big social media company today started by not making a profit for years
English
@TrueSlazac @ordinarytings then a follow up. How do you then explain the amount of venture capital investment into companies with very questionable paths to profitability and circular financing deals? at this point, a large proportion of world's venture capital has been spent on this
English
@stpmtk @ordinarytings I didn’t say “in the next few years”, I have no idea when it could possibly happen because I don’t know what it takes
Also I question the very existence of a bubble
English
@TrueSlazac @ordinarytings so what are the arguments the bubble isn't going to pop? also you should really look into opinions of major AI scientists like LeCunn before formulating opinions on plausibility of superintelligence in the next few years.
English
@ordinarytings I get what you mean but I haven’t heard good arguments about why it couldn’t happen following natural technological advances
To be fair, I haven’t looked for them specifically, but at first glance it feels instinctively plausible
English
@concrete_object ano, propalestinský názor není ukazatel který by indikoval zločinnost, podobně jako proizraelský názor neindikuje progenocidní chování
Čeština
@stpmtk Pozor‼️ Teď nastává ta fáze, kdy budeme předstírat, že propalestinské a radikálně levicové názory podezřelého a jeho aktivita v takto zaměřených uskupeních rozhodně VŮBEC NESOUVISÍ s případnými politicky motivovanými zločiny téže osoby.
Čeština