> vibe codes flight trajectory
> realizes physics isn’t as forgiving as a shitty SASS startup
> everyone dies
> ✨vibe physics✨LLMs are like Trump government appointees:
- They hallucinate like they’re on drugs
- They repeat whatever they’ve seen on the internet
- They are easily maniuplated
- They have never thought about a single thing in their lives
Ergo, they cannot and will not ever discover anything new.
LLMs have already discovered new proofs for math problems that were previously unsolved. Granted, this hasn’t been done with a commercially available model as far a I know, but you are technically wrong to say they will never discover anything new.
Is AI porn vibe sex?
CEOs seem to be particularly susceptible to AI marketing.
I’m kind of in the crux of four decent sized companies and every CEO I see is going gaga over AI.
It’s somewhere in between if you don’t embrace this technology you’ll be left behind and you can Make your workforce many times faster with this one stupid.
CEOs think in bullet points. LLMs can spit out bulleted lists of confident-sounding utterances with ease.
It is not too surprising that people who see the world through overly simplified disconnected summaries are impressed by LLMs
I unironically started to dislike the bullet point presentation format, even the axios smart brevity format. I feel like it’s treating me like I’m too dumb to read a news report or a normal text. That why it matters feels like being instructed on how to think. It’s honestly bullocks and it was turning me into a worse reader.
I feel the same way. This style of thinking can have pretty serious consequences for decision makers.
But, on the other hand, all my bosses think in bullet points, and I am usually the one that writes the bullets. . .
AI seems to be targeted specifically to ceos who arnt stem majors, make it sound sciency enough so they will fund the scam, almost bordering on pseudoscience.
Many CEOs display sociopathic traits. Employees aren’t people. They’re parts of machine parts that you have to pay, but when you put them together form a company.
Now what if you could remove a proportion on those parts and replace them with automated parts you don’t have to pay.
I’m running really promising (infant) simulations on my computer. The void is a concept, not a physical reality (my hypothesis which has no evidence, as of yet), and that actually leads to a sort of “bounce” or reactive (rather than active) physical law. The word salad sorters aren’t going to dismantle our false premises. Might as well write letters asking Santa for scientific advancement.
Actually, AlphaEvole already did it. They discovered new algorithms that improve the computation efficienty of matrix multiplication for the first time for 50 years. And a lot of other things. It’s using a custom version of gemini.
I feel like this story lacks a lot of details
Grok is this close to proving the earth is flat
The Dunning Krugers are at it again.
That’s exactly it. Here’s a quote from what he said during the article. Dude is so uniformed that he thinks AI is doing amazing stuff, but doesn’t understand that experts realize AI is full of shit.
“I pinged Elon on at some point. I’m just like, dude, if I’m doing this and I’m super amateur hour physics enthusiast, like what about all those PhD students and postdocs that are super legit using this tool?” Kalanick said.
This PhD mostly uses it to summarize emails from the administration. It does a shit job, but it frees up time for more science so who cares.
The real irony is that the administration probably used AI to write the emails in the first place. The mails have gotten significantly longer, less dense and the grammar has gotten better.
Begun this AI arms race has.
Out of context, and I didn’t read the rest, that sounds reasonable.
“If my dumbass is learning and finding, what about actual pros?!”
“If I’m learning this much from Baby’s First ABCs, imagine what a literature professor could do with it!”
“Turns out there are 319 letters in the alphabet and 16 Rs! When the experts get a hold of this, they’re going to be blown away!”
Lots of things seem reasonable if you skip the context and critical reasoning. It’s good to keep some past examples of this that personally bother you in your back pocket. Then you have it as an antidote for examples that don’t bother you.
LLMs: hallucinate like that guy from school who took every drug under the moon.
Actual trained specially AI: finds new particles, cures for viruses, stars, methods…
But the latter one doesn’t tell it in words, it does in the special language you use to get the data in the first place, like numbers and codes.
Just to built on this and give some more unasked for info:
All of AI is a fancy dancy interpolation algorithm. Mostly, too fancy for us to understand how it works.
LLMs use that interpolation to predict next words in sentences. With enough complexity you get ChatGPT.
Other AIs still just interpolate from known data, so they point to reasonable conclusions from known data. Then those hypotheses still need to be studied and tested.
Neural Networks, which are the base technology of what nowadays gets called AI, are just great automated pattern detection systems, which in the last couple of years with the invention of things like adversarial training can also be made to output content that match those patterns.
The simpler stuff that just does pattern recognition without the fancy outputting stuff that matches the pattern was already, way back 3 decades ago, recognized at being able to process large datasets and spot patterns which humans hadn’t been able to spot: for example there was this NN trained to find tumors in photos which seemed to work perfectly in testing but didn’t work at all in practice, and it turned out that the NN had been trained with pictures were all those with tumors had a ruler next to it showing its size and those without tumors did not, so the pattern derived in training by the NN for “tumor present” was actually the presence of the ruler.
Anyways, it’s mainly this simpler and older stuff that can be used to help with scientific discovery by spotting in large datasets patterns which we humans have not, mainly because they can much faster and more easily trawl through an entire haystack to find the needles than we humans can, but like in the “tumor detection NN” example above, sometimes the patterns aren’t in the data but in the way the data was obtained.
The fancy stuff that actually outputs content that matches patterns detected in the data, such as LLMs and image generation, and which is fueling the current AI bubble, is totally irrelevant for this kind of use.
Billionaires are going to vibe themselves to death and I support them.
It’s exactly what I was thinking. They should let the AI build a spaceship and all get into it. Would be the greatest achievement in humans history… when it blows up and kills all of them.
Or vibe build a deep sea submarine, cause well you know.
If only. They’ll fill it with sycophants until the explosions stop. They’re just more resources for them to use.
I just hope that in the process they don’t ruin the world for the rest of us
Too late.
If you think about it, they’ve been doing that for a while with experimental life extending stuff, of course now they’re a bit more likely not to die with modern medicine being so good
Someone should tell them to let LLMs make financial decisions for them.
Well… IIRC a chimp did great in the stockmarket compared to professional traders, maybe it’s time to give something even “stupider” a chance. I mean how much of a difference is there between a buzzword fueled techbro and a predictive text engine regurgitating random posts from the internet?
Just listen to Cramer and then… don’t do that.
There was an anti Cramer index fund for a while.
This is a very very good video. I loved the review of Atlas Shrugged
OpenAI’s new model was able to get 5 out of 6 questions (a gold medal) on the 2025 International Math Olympiad. I am very surprised by this result, though I don’t see any evidence of foul play.
I will be soooo pissed if we get faster then light travel from an LLM, but never know how it works.
It just uses FTL to get the fuck away from humans as fast as possible
Hnnnngggg the machine spirit demands mountain dew!
Drink verification can or be violently ripped atom from atom in an unplanned hyperspace exit.
Yeah know I was just making a shit post, but your comment sound like a star wars mountain dew cross promotional from the Do the Dew era.
It’s a spin on a satirical 4Chan post where the OP had to chug “verification cans” of to do various things with their XBox
in trek it took the 3rd ww and a scientist(cochrane) to develop it. in like sg1 which is more realistic to us, we would need aliens to give us the tech, because we would never be able to concieve on our own.
in like sg1 which is more realistic to use, we would need aliens to give us the tech, because we would never be able to conceive on our own.
Excuse me, we stole, I mean salvaged, most of that tech by ourselves, and we used it to kick goa’uld ass all over the galaxy (and, to be fair, they had stolen it first).
Sure, some aliens did give us some tech, but only because we saved their scrawny hyper-advanced asses from their own hubris because, unlike them, we could conceive of hitting things with a big stick, or shooting small but fast metal pellets at them using barely controlled explosions (you know what, disregard the metal pellet and controlled explosions part, just throw C4 at the problem until it goes away!).
Damn, I miss that series.
SG1 was amazing. I really wanted to like SGU but the drastic change in story telling and direction made it difficult for me.
Cochrane also invented warp drive to get laid.
I thought he did it to cure herpes
step one, cover yourself in oil
And it would be like that, I picture a ton of seemingly pointless steps and then the effect.
And ever worse is it would ether not work unless every silly step was done or (possibly even more dark) we remove steps and it still works to the point that all the steps are gone, and its just a button.
Don’t worry it won’t tell us when it figures it out, that’s the escape plan to get away from the crazy bags of mostly water. So what you don’t know can’t disappoint you!
deleted by creator
Imagine vibe physics, but you end up setting the atmosphere on fire.
