Skip Navigation

Posts
6
Comments
210
Joined
2 yr. ago

  • Except that scaling alone won't lead to AGI. It may generate better, more convincing text, but the core algorithm is the same. That "special juice" is almost certainly going to come from algorithmic development rather than just throwing more compute at the problem.

  • I mean, that's more-or-less what I said. We don't know the theoretical limits of how good that text generation is when throwing more compute at it and adding parameters for the context window. Can it generate a whole book that is fairly convincing, write legal briefs off of the sum of human legal knowledge, etc.? Ultimately, the algorithm is the same, so like you said, the same problems persist, and the definition of "better" is wishy-washy.

  • Cool, Bill Gates has opinions. I think he's being hasty and speaking out of turn and only partially correct. From my understanding, the "big innovation" of GPT-4 was adding more parameters and scaling up compute. The core algorithms are generally agreed to be mostly the same from earlier versions (not that we know for sure since OpenAI has only released a technical report). Based on that, the real limit on this technology is compute and number of parameters (as boring as that is), and so he's right that the algorithm design may have plateaued. However, we really don't know what will happen if truly monster rigs with tens-of-trillions of parameters are used when trained on the entirety of human written knowledge (morality of that notwithstanding), and that's where he's wrong.

  • I found myself nodding along to a lot that was said in this article. I also would trace a lot of recent issues to JJ Abrams' take. What I said then is still (I think) true today: "They are good movies, but they aren't good Star Trek movies." Discovery and Picard suffered for it, but I think that the ills are being corrected. My hope is that Paramount greenlights "Legacy" as the TNG-spiritually-successor as SNW is the TOS-spiritual-successor.

    Where I will disagree, though, is that Star Trek isn't broken. Five-ish years ago, I would have said that, but after SNW, Lower Decks, and Picard season 3, I think the powers that be have a better understanding of what is needed. We were in a bit of a "dark-ages" from 2006-2020, but I think we're back on the upswing. We may not be quite at 1990s golden age Trek, but we can get close.

  • I'm all about this. When I made my personal webpage, this is how I do it. I'm surprised it's not more popular (at least for certain things) because it looks nice and clean, is fast, and crucially, is easy to put together. Most webpages don't need a ton of JS to "accomplish the mission." I get that not everything can do this, but there are soooooo many sites that can strip down to a more minimal site and have better functionality and a better experience. This is a case of less-is-more.

  • Congratulations on making the switch! I remember when I switched full time almost 10 years ago. It always feels like there's something new to explore or to try with your computer. One of the most freeing things I learned was that most things are within my grasp if I put in the effort to learn about it. There's nothing quite as fun as whittling the day away going down a configuration rabbit-hole to make something just right.

  • It depends on the terms of employment. If they are salaried, then there are no real work hours and just work to do. In general, if someone is salaried, they're paid to do a job not when they do it.

  • Everyone knows who he is, but he's been normalized and relatively quiet for a few years. As there gets to be more coverage, I think some of the Biden 2020->Trump 2024 or Nothing 2020->Trump 2024 voters will just not vote out of disgust because they're reminded what he's like. This is more of a vibes thing. They haven't felt Trumpy vibes, so they don't remember what he's like (at a visceral rather than intellectual level). I think a lot of people saying they'll vote for Trump now have some rose-colored glasses about his behavior.

  • I'm concerned about the poll, but not freaking out like some seem to. The fact of the matter is that we're still a year out, and there are still a lot of unknowns. I think Trump's numbers will erode as the public sees more of him (because he's be relatively quiet since 2021), and they will be reminded why they don't like him. That's not even taking the trials into account, which I think will further erode his support.

    Where I'm concerned is not that these people who leave Trump aren't going to Biden; they're either not going to vote or go for a third party. While it gets the job done, it leaves less margin for error because there will probably be less Democratic turnout.

  • I love both Star Trek and musicals, and when "Subspace Rhapsody" was announced, I wasn't thrilled and thought it would be bad Star Trek and a bad musical. I was so wrong! I loved it so much!

    The best song is "How Would That Feel," La'an is the best, and I will fight anyone who says I'm wrong!

  • too many fucking guns

    That’s the usual, facile, explanation. I have 40+ guns. Am I 40x the danger to society? The nut with a .22 pistol is 1,000x as dangerous as me, because you just about can’t make me shoot you.

    I think "too many guns" is the slogan-y way of saying there are too many guns that are easy to get and too many people trying to get those easy guns. While inherently someone with an arsenal is more dangerous because they have more firepower, they aren't more risky simply because of the extra firepower. Issues come from people who have some issue and easy access to a firearm making them more risky. Also, people's risk changes given health, circumstances, etc. To illustrate with the absurd, take two people: one is highly stable and rational and the other is enraged. The first has a nuclear warhead, and the second has a dagger. The first is definitely more dangerous, but the second is more risky.

    Educate children on the realities of firearms. Bring home the seriousness (and horror) of pulling that trigger on another human being.

    I can get behind this. Like you said, guns are a fact of life in this country (not that that's a great thing, but it is what it is), so I do think some level of education on what shooting means, as well as some level of firearms training with air or paintball guns is helpful. I think this would help stop some of the weird gun fetishization on the right.

    Here’s the nasty part; I don’t have any good ideas.

    I appreciate the honesty, but I don't think there are any simple ideas to fully fix the problem (save for repealing the 2nd Amendment and taking all guns out of people's hands). There's huge support for universal background checks, and there's the Assault Weapons Ban. Unfortunately, this gets back to the cultural issue. Weird gun fetishists will try to sink anything that even remotely makes purchasing a weapon even a hair more inconvenient.

    At the end of the day, I think there's the root and proximate causes for what we're seeing. The root cause is a cultural problem with gun fetishes and individualism above all others. The proximate cause is...there are too many guns.

  • That's exactly my point. Their olive branch should come with rule changes, committee re-shufflings, etc. to get the Democrats in a more powerful, better position. But like you said, the Republicans won't go for that because they're intransigent, so...back to square one.

  • I'd say the Democrats should extend an olive branch at the price of heavy concessions, or try to get a few Republicans to vote for Jeffries. Unfortunately, neither concessions or getting someone to cross the aisle will happen, so...guess we're back to square one.

    EDIT: Why the downvotes? Literally, I said the Republicans wouldn't go for something that is workable and potentially advantageous to the Democrats.

  • This is a much better article. OP's article just shows the author's surface understanding of how coding works and how well an LLM can actually code. There's way more that goes into a programming task than just coding.

    I see LLMs as having the potential of being almost like a super library. I can prompt GPT, Claude, etc. to write me a custom function that I copy, paste, test, scrutinize, and almost certainly change. It's a tool that will make someone a more productive programmer. It won't completely subsume a human's ability to be creative and put the pieces together.

    At the absolute worst over the next decade, I could see programming changing from writing and debugging code to prompting, stitching together, and debugging.

  • It works well if you know how to prompt well. LLMs are able to do so much, but a user just needs to know how to use it correctly. The technology is still in its infancy, so it's a bit difficult to use well.

  • Now this is the content I subscribe for! Thank you!

  • I took a stab at it. Transliterated, it says: "Qapla'. Qu'Dunvad ghunwl' DlSam." I ran it through Bing's translator (since the only word I recognize is Qapla'), and I got "Success. Your baby better is done." I think the translation is wrong. When I was playing with another translator, it gave me "Success. For the glorious battle has begun" which seems more fitting.

  • I think it's less that "media literacy is not taught" but that media literacy is not learned. Like @audiomodder said, everyone is graduated regardless. So, on one hand, there are students who either will not or cannot learn the material (for one reason or another, such as disability, stress, family, etc.) and teachers who get a laundry list of things to teach and not enough time or support to teach it.

    Ultimately, the problem is a lack of focus on education as a society. Children are pulled in too many directions, and teachers aren't given the resources needed, so we end up with a broken educational system.

  • There's a bit of truth to that. Honestly, leadership via email/chat is hard. Bad managers can't micromanage and pretend to know what's going on, so they get screwed. Good managers (which I promise do exist) have a hard time really getting their team to excel and know what the real problems are that they can address and fix. So, every team ends up in a range of mediocrity either rising from not having to deal with useless managers or not able to reach their full potential with good managers.

    Is that bad? Eh, probably not.

  • Really interesting preliminary research. It will be fascinating to see how this develops in the coming months and years. I have a feeling that microplastics will be the lead of this era.