VoxelVexillologist
🇺🇸 Multidimensional Radical Centrist
No bio...
User ID: 64
VoxelVexillologist
🇺🇸 Multidimensional Radical Centrist
2d ago
It does seem interesting that the high-end model guardrails have converged on IT security and biology as their two focuses. I don't have a third branch of "existential dangers easily triggered by AI" to suggest, but if I did I wouldn't be posting about it.
VoxelVexillologist
🇺🇸 Multidimensional Radical Centrist
2d ago
I see the logic in that, but if you accept that I'm not sure where you'd draw a bright line in technology from checks notes agriculture to AI. If you believe technology inevitably puts us on the path to the Great Filter, "retvrn to hunter gatherer" isn't like, obviously wrong.
VoxelVexillologist
🇺🇸 Multidimensional Radical Centrist
2d ago
what if it turns out even very "hard" or concrete fields like electrical engineering or chemistry are built on a bunch of stupidity that has just enough facade to be convincing to the in-field expert and certainly more than enough to the layman?
Your phone works, doesn't it? Everything from the 5G radio protocols through the processors and OLED screens here isn't trivial and relies on a lot of in-field expertise that can demonstrably be taught, but isn't obvious to the layman.
If it's a facade, it's one that is at least very functional and works reliably day in and day out. Sociology isn't demonstrating it's functional correctness billions of times per second on each of billions of devices.
VoxelVexillologist
🇺🇸 Multidimensional Radical Centrist
2d ago
The complaints I recall hearing from the less-theoretical physicists is that the string theorists were taking up a large share of the available funding effectively navel-gazing daydreaming about the pretty symmetries of strings without actually producing testable hypotheses even at "LHC" scales of "testable". Those folks are used to expensive tests, but the string theory results needed like 3-4 or more orders of magnitude in scale.
VoxelVexillologist
🇺🇸 Multidimensional Radical Centrist
2d ago
I'm curious why our self-declared experts thought "alignment" was a solvable problem. It's quite possible that the same logic and operation will be drastically different in ethics in ways that aren't discernable to the low-level workers. The same designs and production lines making potentially-civilization-ending Titan II ICBMs were used to make the launchers for the peaceful Gemini missions. Sometimes it's clear from the context and we feel comfortable assigning blame (where did all those box cars of people go?), but there are plenty of historical examples of humans not knowing the moral valence of the larger efforts they worked on.
As much as I find the idea of Asimov's laws of robotics comforting, the stories he wrote are mostly about the inadequacies of those rules.
VoxelVexillologist
🇺🇸 Multidimensional Radical Centrist
2d ago
In physics, you usually can turn a model into predictions which will eventually be measurable,
For a while, the String Theory folks got a lot of funding, but never seemed to deliver practically testable predictions. I'm not sure this is as true in physics as you'd like it to be, but it is certainly moreso than the humanities.
Sam himself seems spooked by the HF hack (as he should be).
I'm not sure if I believe this is his earnest take, or an act. Really for the whole company. These are the folks that have been telling us how dangerous this AI thing could be. And they are the ones that seemingly took off the guardrails and put their model behind only a proxy with open Internet access. What did they think would happen? I'm not sure if I believe they are that stupid. There are well-known ways to make different levels of air-gapped systems!
The whole thing reads like Ian Malcolm (played by a shirtless Jeff Goldblum) was found to have, while decrying the dangers of Jurassic Park, an incident at his own live dinosaur theme park, but it turns out that they built their Velociraptor paddock relying only a single strand of barbed wire for containment. Quite the conflict of interest, to say the least.
VoxelVexillologist
🇺🇸 Multidimensional Radical Centrist
2d ago
the increase of salt concentrations in the soil of the Fertile Crescent due to poor irrigation practices also doomed that region to relative irrelevance in the modern era.
Looks at map. While you're probably right in terms of its agricultural output, I'm not sure I'd describe a region including modern Israel, Lebanon, Syria, Iraq, and Kuwait to be "irrelevant" in the modern era. It definitely punches well above its geographic-area or population weight in news headlines.
VoxelVexillologist
🇺🇸 Multidimensional Radical Centrist
3d ago
Old and new reddit do that inconsistently, even. I found using the URL percent escape code for close paren %29 works most reliably across implementations, but YMMV.
VoxelVexillologist
🇺🇸 Multidimensional Radical Centrist
3d ago
I could see that angle, but in the Star Wars universe I assume it'd be Empire-skepticism: "fought in the Clone Wars, now just getting by with the Imperial overlords with some chafing".
That "today's Hollywood might decide to stan the Empire (literal Stormtroopers carrying kit-bashed Nazi MG 34s) over a potentially-problematic weak lost cause allegory" seems to succinctly describe what is wrong with the political zeitgeist today, IMO.
Currently on Seveneves. So far my least favorite Stephenson book, but I guess I'm comparing it to Diamond Age, Snow Crash, and Cryptonomicron, probably two of which are in my "top 10 books I've read" list. It's okay, about a third of the way through. Feels very much "of its time" (2008 2015): drops Internet references (Reddit/YouTube), and two of the characters seem obvious NDT and Elon Musk ("when he was still cool") stand-ins. No questions asked international cooperation. It's long (800+ pages). Not exactly a slog, but I'm tempted to set it aside for something more interesting. Thinking As I Lay Dying, since I've found I like "in the style of Faulkner" AI memes.
Previous was True Grit, which is a short read I'd recommend. Might watch the movies at some point.
VoxelVexillologist
🇺🇸 Multidimensional Radical Centrist
4d ago
Same recipe, fresh ingredients.
Given that Disney owns both now, I've been thinking a spiritual successor to Firefly would fit well in the Star Wars universe. But I wouldn't trust the company blindly to produce a good show these days.
Andor was good, though. Same idea; new characters and setting.
VoxelVexillologist
🇺🇸 Multidimensional Radical Centrist
4d ago
They even managed a popular spin off TV show from it!
VoxelVexillologist
🇺🇸 Multidimensional Radical Centrist
6d ago
It does seem under this definition like the parties aren't really as conservative/progressive aligned as the messaging would have you think. Under "don't do anything new", California is arguably a conservative state when it clamps down on building housing, sets strong environmental laws discouraging industry, and blocking data center construction. Alternatively, Texas's libertine rules encouraging solar build out, natural gas, and encouraging industry are progressive. Don't tell Gavin and Greg.
But that seems part of both party's ongoing civil wars: blue team seems to have inherited a lot of conservative (lacking a better term) seniors for whom that means "hands off my Medicare" and a general preference for maintaining late 20th century neoliberalism, which clashes with the DSA types wanting radical changes. Red team has a similar clash between traditionally-labelled conservatives and "you can just build things", although I'm not sure if that fight is still ongoing these days.
It seems quite plausible to me that Fable was expected to be profitable, and the promised multibillion-dollar capital moat to afford training compute has not held back competitors as much as expected. And Sam, at least, has to make good on his promise to buy all the RAM, cynically I'd say to keep prices high for competitors.
At some philosophical level, it seems worth noting that the real Claude Shannon was effectively "trained" with a much smaller pile of books and a couple decades of a few dozen watts. I'm not convinced training will always be so expensive.
I'm also reminded of the National Ignition Facility, which uses huge lasers for, uh, "fusion research". It was designed at a time when lasers were very inefficient, and we can do much better these days. However, it wasn't built for economical energy production (under 1% laser efficiency) and does make the pulses it needs to, so they have no reason to replace it. If it were a commercial effort for energy production, they'd have lost their shirts to smaller startups with cheap diode lasers (50%+ efficient). Sometimes moats don't work like you planned.
More options
Context Copy link