FaceDeer

FaceDeer, 2 months ago (edited 2 months ago)

Most fun I can recall recently was pointing out how Meta is actually one of the main driving forces behind the availability and development of open-source large language models. Meta's pytorch framework is one of the foundation pieces of many LLMs industry-wide. Meta has released a bunch of major open-source libraries and frameworks. Their open LLaMA model weights are the starting point for many fine-tuned models floating around out there.

But no, Mark Zuckerberg is an evil lizard man, so can't mention anything nice he might be responsible for.

Doesn't help that AI is a hot-button topic in its own right.

FaceDeer, 2 months ago

More likely you're getting downvotes for the fallacious "I'm obviously being the logical one, everyone disagreeing with me must be illogical" form your comment takes.

And deservedly so.

FaceDeer, 2 months ago

Saying something that is insufficiently negative about Elon Musk. Or Mark Zuckerberg, he's another one on the "villain" list.

FaceDeer, 2 months ago (edited 2 months ago)

They are where people tend to live.

This winter my home city had a power supply crisis. It was night time (I live in a high latitude so nights last a long time during winter) which meant no solar, and it was -30C, which meant the wind turbines all shut down (they can't operate when it's below -30C). The whole province was short of power, only the coal and natural gas plants were keeping the lights on. We dodged rolling blackouts but it was a close thing. Lots of people live here.

Bring down the nuclear level

Which is perfectly fine. Nuclear power plants can change how much power they're putting out. It's not "economic waste", the term is "load-following power plant" and it's routine for nuclear power plants.

FaceDeer, 2 months ago (edited 2 months ago)

It weakens your claim that it's not hyperbole, since the meaning of words like that are consensus-based.

Do you seriously think it would make sense for a 68-year-old man to claim that he's "nearly a century old"? That wouldn't be misleading?

FaceDeer, 2 months ago

So you agree that hyperbole is uncalled for?

FaceDeer, 2 months ago

I'm not the only person calling you out on it.

FaceDeer, 2 months ago

Massive downside to that: photos will exist of Trump shirtless.

FaceDeer, 2 months ago

I can't recall the last time I pirated anything executable (games and other software). There are legitimate free options for everything I've wanted, and executable code is just too risky.

FaceDeer, 2 months ago

This article is projecting 76 years forward, that's not the near future any more.

FaceDeer, 2 months ago

76 years ago was 13 years before Yuri Gagarin would become the first human in space. It was 4 years after the V2 rocket became the first artificial object to enter space. This is plenty of time for multiple technological revolutions to happen. We're already on the verge of one with fully reusable superheavy lift rockets, most people don't grasp just how big a change will come from having that sort of cheap bulk cargo access to space.

it'd probably be a few hundred years until we could actually make nation-sized space colonies

There's no need to make nation-sized space colonies, just make lots of smaller ones.

There were a few interesting astrophysics papers estimating that near-lightspeed and FTL travel tech is like 8000 years away lol.

I would like to see those papers. Making technological estimates on that scale, especially for something like FTL that has no physics backing it at all, is highly dubious.

"Future technology" can't solve all of our problems. It's not magic.

There's no need for magic, this is really just a question of economics.

FaceDeer, 2 months ago

Kessler syndrome is only a threat to satellites that are orbiting within the debris, it's not really a danger if you're only passing through (as a ballistic missile would).

FaceDeer, 2 months ago

Those steeply elliptical orbits would probably deorbit even quicker since a random impulse that boosts the apogee is likely to lower the perigee even more.

FaceDeer, 2 months ago

It's okay, Starlink is in a low enough orbit that it's basically Kessler-proof.

FaceDeer, 2 months ago

Be specific. Which ones?

No. You are the one who said "Using your metaphor the thing you're proposing to "treat the symptoms" has side effects which worsen the disease thus causing more real damage and worsening symptoms."

I then asked you to be specific. You tell me which ones have side effects that "worsen the disease." You don't get to Uno-reverse at me until you answer the question first.

FaceDeer, 2 months ago

But as you said, stopping emissions won't avoid decades of worsening conditions. I think actually stopping those decades of worsening conditions is more important than a hypothetical "moral hazard" concern.

Frankly, this argument always bothered me. When someone is sick you try to treat both the underlying cause and the symptoms. It would be morally bankrupt and downright ridiculous to say "let the patient suffer, it's the only way he'll learn." Especially if the symptoms themselves could be fatal. And especially when the people suffering aren't the ones who actually "need to learn." When millions of people are starving to death in third-world nations or drowning when their overloaded refugee ships are turned away from wealthy ports, will you look them in the eye and tell them it's necessary because otherwise oil company executives might not be as motivated to reduce emissions?

FaceDeer, 2 months ago

It also isn't going to avoid decades of worsening conditions, but there isn't a solution for that.

Various geoengineering techniques are solutions for that. We should be studying those in greater detail.

FaceDeer, 2 months ago

Using your metaphor the thing you're proposing to "treat the symptoms" has side effects which worsen the disease thus causing more real damage and worsening symptoms.

What side effects, specifically? Some approaches to geoengineering may have negative side effects, but others don't appear to. There's no guarantee that an approach without side effects won't be found.

You aren't saving "millions of people from starving to death"

Yes, you are. Climate change would cause famine, ameliorating the effects of climate change would prevent that famine.

This whole comment is exactly the kind of argument that I'm objecting to. You've got some sort of a priori conviction that "no, geoengineering must make the situation work somehow" and therefore it's not worth studying. If it's not studied how can you possibly know?

FaceDeer, 2 months ago

Ah, tucked away down at the bottom. mBin hides the full content of long responses and I was admittedly getting quite frustrated talking with you after you responded to my "be specific" request with a "be specific" of your own.

The article appears to have two main criticisms:

• Kelp might outcompete phytoplankton
• Kelp might not actually grow well enough to work

Those are basically "it might work too well" and "it might not work." I don't see anything in there that would make climate change worse.

Personally, I'm not terribly interested in the carbon sequestration approaches. They seem unlikely to be able to be scaled well enough to have an impact in an economically realistic way. Solar radiation modification is IMO the most likely class of approaches to geoengineering to help.

FaceDeer, 2 months ago

Do we count those as negative numbers when the implants help them recover?

FaceDeer, 2 months ago

It's exactly that "new loud wave of complainers" I'm talking about.

I've been in computing and specifically game programming for a long time now, almost two decades, and I can't recall ever having someone barge in on a discussion of game AI with "that's not actually AI because it's not as smart as a human!" If someone privately thought that they at least had the sense not to disrupt a conversation with an irrelevant semantic nitpick that wasn't going to contribute anything.

FaceDeer, 2 months ago (edited 2 months ago)

Joke's on you, I skimmed by it without engaging for more than three sec... shit.

FaceDeer, 2 months ago

There's an assumption in the comments that this is Lemmy-specific, so I figured I should also mention a tool I used recently when copying subscriptions from a kbin instance to an mbin instance.

FaceDeer, 2 months ago

I wonder where the goalposts will shift to next.

FaceDeer, 2 months ago (edited 2 months ago)

No, the goal posts of "AI is evil and should be fought until <insert new criteria here> are resolved."

The question of whether training an AI even violates copyright in the first place is still unanswered, BTW, the various court cases addressing it are still in progress. This current target is about "ethics", which are vague enough that anyone can claim that they're being violated without having to go to the hassle of proving it.