eubanksengineeringresearch.com
Professional engineer and permaculture designer fascinated by the thermodynamics & metabolics of infrastructure and the built environment.
Humanity's future lies in cooperation with natural systems. Stop fighting with the planet - it is bigger than you!
6,996 posts
1,800 followers
1,082 following
Regular Contributor
Active Commenter
comment in response to
post
Which is why he’s on TV, natch.
Good old Governor Hairgel. You can always count on him. That is to say, he’s predictable.
comment in response to
post
Stopped clocks, etc
comment in response to
post
Trump will be funny later, after this is over.
Newsome is absurd now.
comment in response to
post
In other news, water is wet.
I also heard that the Pope is Catholic.
The new one too, despite what the conservatives would have you believe.
😉
comment in response to
post
And some plants just naturally have leaves that color.
comment in response to
post
I suppose there are other more believable explanations for LAPD including “too scared” and “can’t be arsed”.
comment in response to
post
That’s one of those either/or situations that’s like trying to approach two of the same pole of a magnet. You consider one and go “surely not” and then face the other possibility and find it equally implausible.
comment in response to
post
I thought that this was the one actually useful thing that Bitcoin et al were good for.
comment in response to
post
Trump should just nationalize SpaceX as a strategic asset.
comment in response to
post
As far as I can tell, Blue Origin hired a bunch of OldSpace execs and has basically been recapitulating their playbook. They’ve barely launched anything despite an over decade and $Bs spent.
Musk is an asshole but SpaceX’s R&D culture is the correct one for rocketry, or at least it used to be.
comment in response to
post
Anything that the brain can entrain on… for me it’s some sounds, some of the time. For some people it’s apparently all of the time.
See also sensory integration disorder/sensory processing disorder.
comment in response to
post
Played like a fiddle…
comment in response to
post
Is absolutely on brand, you mean.
comment in response to
post
It’s worth noting that there is one thing above all else that is hardest of all to decarbonize - in fact, that is IMPOSIBLE to decarbonize regardless of technology, process or any other change one might imagine:
Billionaires.
We have to get rid of billionaires. They are incompatible with survival.
comment in response to
post
Other things that are (much!) harder to decarbonize than AI:
- heavy industry (e.g. anything involving steel)
- air travel
- food production
- construction
Our civilization’s relationship with the environment is completely broken, but AI is barely a footnote in that story.
comment in response to
post
I hate to break it to you, but that was already done for by… *looks around* (at everything, but especially at cars)
AI is infinitely easier to decarbonize than personal automobility, and the huge resource footprint of LLMs reflects economic and research priorities, not technology fundamentals.
comment in response to
post
What price America’s soft power and reputation as a reliable global partner?
To say nothing of democracy and the rule of law...
comment in response to
post
Ok I posted my other reply and then I saw that you had data science in your profile. LOL. No offense intended. I think you get the distinction I’m pointing at.
comment in response to
post
Oh I am well aware lol. Ecologically illiteracy runs rampant in STEM but nowhere more so than in software “engineers” and data “scientists”.
But the AI people should at least understand the limitations of the AIs ability to “think” beyond the scope of their own training data, but they don’t seem to
comment in response to
post
It would be really unfortunate if the people who care about the ethical use of technology to reject companies that adopt the technology contientiously, because that will ensure that the ONLY companies that survive will be those that adopt it ruthlessly and exploitatively.
comment in response to
post
It would be really unfortunate if the people who care about the ethical use of technology to reject companies that adopt the technology contientiously, because that will ensure that the ONLY companies that survive will be those that adopt it ruthlessly and exploitatively.
comment in response to
post
2. Any knowledge/information company that refuses to adopt the broader set of machine learning technology at all, in any way, will be crushed by competitors who use it aggressively.
The ethical ones that survive will be the ones who adopt it defensively. That _seems_ to be what is happening here.
comment in response to
post
It’s a moot point for me. It’s been years since I’ve had the privledge of time to read for pleasure.
But two things I can say for sure:
1. The small models are different from the generative ones in every way that matters. Kobo's recognition of this seems like a basis for more trust, if anything.
comment in response to
post
Oh, no, you’re understating the case if anything.
An LLM is a cognitive mirror. Which means that for anyone with any shred of narcissism (i.e. anyone with a pulse and most people recently dead) a chatbot has the potential to become the ultimate ego trap.
comment in response to
post
It’s in all forms of error correction just for starters. All image and audio recognition. All kinds of embedded systems.
That’s been true for years, long before ChatGPT etc.
comment in response to
post
But what’s even your objection? That they use ML? That they use computers at all?
I mean, if you want to avoid ML algorithms I’m sorry to say but you’re going to be living a pretty primitive lifestyle.
ML is literally EVERYWHERE.
comment in response to
post
Yeah fair enough. Aside from the CEOs, though, it seems like a lot of the people making crazy AI pronouncements in public are technical people who really ought to know better.
I suppose they’re just compartmentalizing and “talking their book”. Bunch of assholes though.
comment in response to
post
The AI researchers aren’t that. They _can’t_ be that, most of them, and do what they do.
And yet most of them are either lying or are failing to apply what they already know about their own creation to their statements about it.
comment in response to
post
Also a product of doing all their work in a world without inertia, without friction, and where any mistake can be corrected by a restore from backup.
comment in response to
post
Probably true for 99%. I am increasingly interested in the corner case however.
comment in response to
post
Proving that they fundamentally don’t understand this technology, either.
comment in response to
post
The definition the TSHTF is that we all lose everything, to first order approximation.
They have more to lose, relatively, but we all have far more to lose than most of us realize.
comment in response to
post
They’re not using Large Language Models. They’re using small, special-purpose models. These don’t cost a mint to build or take Hoover Dam to run. You can run some of them on your home PC.
All LLMs are ML but not all ML are LLMs.
comment in response to
post
Huh. Didn’t actually know that bit.
Of course he doesn’t have any real influence himself, so it doesn’t matter what he tells people to do.
His only asset is the ability to generate intellectual smoke thick enough to cover for dumb rich people who want to pretend that they’re smart.
comment in response to
post
Granted.
But that’s not to say that voting for Biden as an actual dead, taxidermied corpse would not still have been the most reasonable option of the ones presented to us.
comment in response to
post
The most insidious double standards are the ones that EVERYONE applies because you can’t expect a toddler and an adult to behave similarly.
The media’s preference for an in-boring president just made it worse.
comment in response to
post
But I’m still not sure what your point is.
Voters are irrational? Yes. Clearly.
“Low information voters” are marks? Also clearly true.
What I don’t understand is you seem to be trying to present the voters’ decision as if it was rational or informed, when it clearly was not.
comment in response to
post
So you had two candidate who were both totally unfit, for entirely different reasons, but one of those unfitness scenarios was objectively much more dangerous than the other.
One of those candidates was also clearly preferred by the media ecosystem.
comment in response to
post
Meaning that no one knows where he lives?