Not high on the list of thought crimes, but a particular ick for me:
Also: 100% of all the Bash code you should author ever again
Why the bash hate?
zogwarg
joined 2 years ago
Oh no! I wasted my time on Troll. Typical.
Hard disagree, as much as I loathe JK Rowling's politcal ideas, and the at-times unecessary cruelty found in the HP novels, it still shaped a large part of the imaginary world of a generation. As beautiful as bird songs are (who the hell refers to birdsong as "output"), this simply cannot be compared.
Yes commercial for-profit shareholder-driven lackadaisical "art" is already an insult to life and creativity, but a fully-or-mostly automated slop machine is an infinitely worse one.
Even in the sloppiest of arts I have watched, the humanity still shines through, people still made choice, even subjected to crazy uninispired didacts from above, the hands that fashion books, movies, music, video-games, tv-shows still have—must have—room to bring a given vision together.
I think people DO care.
I don't know exactly what you wanted to say, if you wanted to express despair, cynisism, nihilishm or something else, but I would encourage you not to give up hope with humanity, people aren't that stupid, people aren't that void of meaning.
The standout monuments of stupidity—and/or monstrosity—in McCarthy's response for me are.
- Calling JW a failed computer scientist for failing to see that computers and clockwork are different, when really there is no computation a computer can make that Turing Complete clockwork couldn't be able to replicate.
- Essentially saying that by analogy, where religion should not stand in the way of science, so should morals not stand in the way of science?!?!?! (I mean really? WTF)
Rekindled a desire to maybe try my own blog ^^.
I think beyond "Keeping up appearances" it's also the stereotype of fascists—and by extension LLM lovers—having trouble (or pretending to) distinguishing signifying and signified.
Seriously though, I can i trust dotnet ever again?
Infinite-garbage-maze does seem more appealing than "proof-of-work" (the crypto parentage is yuckish enough ^^) as a countermeasure, though I would understand if some would not feel confortable with direct sabotage—say for example a UN organization.
I feel the C-SUITE executives are pushing the AI way harder than they ever pushed crypto though, since they never understood the tech beyond a speculative asset, but the idea of replacing work-hours by AI-automation has been sold HARD to them.
I guess the type of lawyer that does this would be the same that would offload research to paralegals, without properly valuing that as real work, and somehow believe it can be substituted by AI, maybe they never engage their braincells, and just view lawyering as a performative dance to appease the legal gods?
This is beyond horrifying:
I don't know to decide wether I should be glad this wasn't show to a jury, or sad we don't get an obvious mistrial setting some kind of precedent against this kind of demented ventriquolism act, indirectly asking for maximum sentencing through what should be completely inadmissible character testimony.
Does anyone here know how 'appeals on sentencing' vs 'appeals on verdicts', obviously judges should have some leeway, but do they have enough leeway to say (In court) that they were moved for example by what a spirit medium said or whatnot, is there some jurisprudence there?
I can only hope that the video played an insignificant role in the judges decision, and it was some deranged—post hoc—emotional—waxing 'poetic' moment for the judge.
Yuck.
It's also such a bad description, since from their own post, the Bot+LLM they where using was almost certainly feeding itself data found by a search engine.
That's like saying, no I didn't give the amoral PI any private information, I merely gave them a name to investigate!
EDIT: Also lol at this part of the original disclaimer:
An expert in LLMs who has been working in the field since the 1990s reviewed our process.
1
AGI Sparklings proponents rejoice! Finding a literal map(*) means LLMs have a world model.
(awful.systems)
Transcribed:
Max Tegmark (@tegmark):
No, LLM's aren't mere stochastic parrots: Llama-2 contains a detailed model of the world, quite literally! We even discover a "longitude neuron"Wes Gurnee (@wesg52):
Do language models have an internal world model? A sense of time? At multiple spatiotemporal scales?
In a new paper with @tegmark we provide evidence that they do by finding a literal map of the world inside the activations of Llama-2! [image with colorful dots on a map]
With this dastardly deliberate simplification of what it means to have a world model, we've been struck a mortal blow in our skepticism towards LLMs; we have no choice but to convert surely!
(*) Asterisk:
Not an actual literal map, what they really mean to say is that they've trained "linear probes" (it's own mini-model) on the activation layers, for a bunch of inputs, and minimizing loss for latitude and longitude (and/or time, blah blah).
And yes from the activations you can get a fuzzy distribution of lat,long on a map, and yes they've been able to isolated individual "neurons" that seem to correlate in activation with latitude and longitude. (frankly not being able to find one would have been surprising to me, this doesn't mean LLM's aren't just big statistical machines, in this case being trained with data containing literal lat,long tuples for cities in particular)
It's a neat visualization and result but it is sort of comically missing the point
Bonus sneers from @emilymbender:
- You know what's most striking about this graphic? It's not that mentions of people/cities/etc from different continents cluster together in terms of word co-occurrences. It's just how sparse the data from the Global South are. -- Also, no, that's not what "world model" means if you're talking about the relevance of world models to language understanding. (source)
- "We can overlay it on a map" != "world model" (source)
view more: next ›
Special bootlicking points:
Source: xcancel.com
For reference Kurt is the CEO of the company that the author works for: https://archive.md/Z2xvg