Yes, exactly.
I mostly don't understand why they didn't wait a bit longer to have a new Zelda or a new Kart or a new Smash at the same time. I really like Mario Kart World. The open world is fun, but I think there are also a lot of missed potentials there. Competitive time challenges and so forth..
I'm hoping for some big updates down the line.
The cartridge issue is really annoying, but aren't most of the other consoles like that now? I haven't touched a Playstation or an Xbox in several generations.
What's the deal with Switch 2 hate? Lackluster upgrade?
My kids got one. Mario Kart is really solid. A number of older games have free upgrades which make them nicer to play. It doesn't seem especially compelling to me (I imagine we'll be waiting some time for new Mario, new Zelda, new...really anything), but the hardware seems pretty decent?
Ultimately, we'll see.
In the meanwhile, I'm investing in energy (VDE) and chipmakers (NVDA, AMD, etc). My best guess is that they'll be growing for some time yet. Chipmakers and energy are needed no matter which AI companies or technologies succeed.
I should add that very little of this is anything original to me or my insights. Here’s a really good article from 10 years ago:
https://waitbutwhy.com/2015/01/artificial-intelligence-revolution-1.html
It’s possible that LLMs are a dead-end technology—that we’ve already maxed out the possibilities. While ChatGPT 3.5 was a huge step over what came before, 4o is still sometimes considered the best ChatGPT model, and there’s a decent amount of dissatisfaction with the 5.x series.
There’s a possibility that with the proliferation of AI slop—specifically, LLM-generated text—on the Internet, that future attempts to train LLM models on Internet data are going to see major regressions.
It’s possible that context windows provide a hard limit and that increases in the context window don’t “scale” in terms of functionality.
It’s possible that to scale models further we need exponential increases in CPU and memory capacity beyond what we can effectively create.
It’s also possible that we can get 99% of the way to AGI, but that last step is really, really hard.
I’m not saying any of that is likely, but I also think AGI is not (yet) a sure thing anytime soon.
My own personal bet would be that, outside of an entirely non-LLM technology, future models will run dozens of LLMs in parallel (a logic model, a math model, a physics model, a history model, a science model, an art model, etc. - some of these may not even by LLMs.) and then another layer of models judges, sorts, selects, adjusts, those outputs. Multiple levels of this. Context window is still a big issue.
The more computing capacity that is available, the more options there are.
I agree LLMs aren't going to be AGI's but they're surprisingly good at pretending to be intelligent. I think AGI could happen within 20 years. Others in the field think it will be under 10 years.
From my own mostly uninformed perspective (I took one AI class ~20 years ago when Go was considered an impossible challenge and our final project involved writing a very simple neural net program to decode single handwritten alphabet letters), it seems to me that the improvement in AI performance largely tracck improvements in CPU and memory capacity. The techniques seem to just largely progress...plateau...progress..plateau while computing capacities go exponential.
I could absolutely see a human-level AGI in 10-20 years, but I could equally see it not happening in 100 years.
I'm interested in the number of thirst posts about the guy...
I'm not remotely a conservative. I'm an anti-leftist.
AIs that will eventually outperform humans in every way
Possible or probable on a long enough timespan, but I don't see LLM technology leading to human or super human level AGI soon. I will admit there are some very compelling arguments to be made along the lines of Kurzweil's singularity, but that's going to happen whether some shitty D&D sub bans AI or not.
Bans don't work. Drug bans don't work. Morality (sex) bans don't work. Drinking bans don't work. Gun bans don't work. AI bans will not work either.
It's wilful ignorance to overlook the enormous downsides far outweighing the benefits of most modern technologies
2500 years ago, Socrates bemoaned the impact of writing on reasoning, memory, and discourse. 2000 years later, the Ottoman Empire found itself getting left behind in part due to their reluctance at adopting the printing press which was again, in large part, exactly along the same lines you mention--"Think of the scribes--jobs will be lost. Humanity will be lost. Art will be lost." Etc. Shall we get rid of writing, and unlock the true potential of the untainted human mind?
What I take is a pragmatic utilitarian position. LLM technology exists. Trillions of dollars are being invested in AI research, AI hardware, utilities, energy, etc. As long as that is the case, I plan to take advantage of the tech.
While we are getting rid of technology that significantly takes us away from the natural state of things, let's start with cars. Cars allow the weak to move freely, competing with the strongest men. Cars (and airplanes) allow people from poor countries to easily move elsewhere. Cars allow us to live physically far apart from each other, destroying communities and relationships that should be much closer and more tightly knit.
And I say that as someone who was on LUE, SA and 4chan as an OG 25 years ago.
The Internet of the late 1990s and early 2000s was so cool. Social media fucked it all up.
20 years ago is literally the golden age or even 20-25 years ago.
Interesting. I had an old friend of mine (who has been playing since the 90s) tell me that now is the golden age. More players, more things like Zoom to make things easier, more media and acceptance, etc.
If we had an actual mod, I would suggest that should be your flair!
So you're taking the same philosophical position that, for instance, the average leftist takes on guns. Guns kill people, guns are bad, therefore, ban guns.
AI hurts people, AI is bad, therefore, ban AI.
I take a different position. Guns are tools. AI is a tool. If someone is using AI to do something cool, why not allow? Sure, ban shlock and low quality posts, just like many forums ban memes and low effort posts.
I stopped playing much WoW a few years ago, though I have played some of the Remixes and classic. Blizzard is majorly pozzed, and I just couldn't take it anymore.
My oldest son liked tanking--running straight into danger and all that.
My non-gamer (other than WoW) wife was, of course, the healer.
We stopped at a WoW dungeon team!
I left to go cut our grass during one childbirth-cut the grass, cleaned the kitchen, got everything tidy. Still had ~8 hours of labor left after that.
Congrats on the baby!
Merry Christmas all!
My family started doing a goose a few years back. It's...interesting.
It's amazing that the government is still making these mistakes (assuming it's not deliberate).
This kind of error is as old as PDFs. The text is there, it's just covered with a black blob.
Policy at CIA, NSA, etc. was to NEVER use PDF or electronic features for redacting text. While it can be done correctly, it's easy to mess up. Physical masking or physical removal (better) of the redacted text, then rescanning documents into fresh electronic format. I would think with today's technology, even physical masking is defeatable in many situations.
Couldn’t care less about that part. You weren’t nasty and mean and bitter back then..
Edit: Well, maybe bitter, but you certainly weren’t obscene.
I liked the old Imp better.
I thought it was an old print term as well (typesetting and moveable type is kind of a hobby of mine), but it terms out "lede" only goes back to the 1980s..
"Mind your ps and qs," "uppercase" and "lowercase" (which case the letters were physically stored in), "against the grain," "make a good impression," and a bunch of other common expressions DO go back centuries or more and come from typesetting and printing.
It is a mystery.
Got it.
How singularly bad are Jews compared to Muslims or Mormons? Polytheist Hindus?
Now you're just arguing in bad faith. You asked for the "historical reasoning" and I, and other posters, have given you several historical explanations and examples. If you want to have a conversation about Christian and Jewish theology, prophetic fulfillment, theodicy, polemics, etc., there have been millions of words written on those topics over the last 2000 years. Read some books.
So you don’t roast forever in hell
I'll give you that one--fear of punishment is part of the decision to convert, or not to convert, for some people.
Because he fulfilled the prophecy they claimed to support.
Clearly they didn't, and don't, agree.
No?
Yes. Moses and the covenant. The Jews were selected by god for a reason (this is their belief, not mine). Now Jesus says, if you follow me, there is no covenant. All men are equally the children of God.
If you are a well off Jew, why would you give up your honored position to become one of everyone else? Clearly some did, but many did not.
Across the world when looking at religious conversion, the groups most likely to convert are the poor, those without social standing, those without power.
Poor Christians in Egypt and the Middle East were universally the most likely to convert. For centuries in Egypt and the Ottoman Empire, Christians actually formed a wealthy minority. The poor Christians who couldn't pay the jizya (Islamic tax on non-Muslims) could just convert and no longer pay the tax.
Conversion and religion is NOT solely a spiritual decision.
I absolutely share many of those complaints. DS Mario Kart was maybe my favorite, and I was really obsessive about 3 stars and the challenges. That's what I hope they add to MKW. I wish there was more of that.
My kids love the open world. Online play is good too. Knockout Tours are a lot of fun.
Like I said about the Switch2 more generally, it does feel a bit unfinished.