Mikina @ Mikina @programming.dev

Posts

5

Comments

482

Joined

2 yr. ago

Thank you, it was an interesting read.

Unfortunately, as I was looking more into it, I've stumbled upon a paper that points out some key problems with the proof. I haven't looked into it more and tbh my expertise in formal math ends at vague memories from CS degree almost 10 years ago, but the points do seem to make sense.

https://arxiv.org/html/2411.06498v1

While I haven't read the paper, the comment's explanation seems to make sense. It supposedly contains a mathematical proof that making AGI from a finite dataset is a NP-hard problem. I have to read it and parse out the reasoning, if true, it would make for a great argument in cases like these.

https://lemmy.world/comment/14174326

If you look at it from this perspective, it sounds way more obvious. I like this PoV.

Lol. We're as far away from getting to AGI as we were before the whole LLM craze. It's just glorified statistical text prediction, no matter how much data you throw at it, it will still just guess what's the next most likely letter/token based on what's before it, that can't even get it's facts straith without bullshitting.

If we ever get it, it won't be through LLMs.

I hope someone will finally mathematically prove that it's impossible with current algorithms, so we can finally be done with this bullshiting.

Kagi can do this by default, without having to hope that a random extension doesn't one day suddenly decides to update into a infostealer. In general, apart from the few super popular ones, installing a random extension that can do the random niche thing you need is a pretty big risk.

I speak from experience, few years ago the developer of Nano Defender, which was at the time better at avoiding anti-adblock scripts, decided to sell/handout the extension to someone, who turned it into a cookie/info stealer, which got through automatic update and started wreaking havoc on everything I had logged in. Since then, I avoid extensions as much as possible.

As for Reddit having the answers - nah, never had an issue with finding what I need without reddit, for the last year I stopped using it, and in the few cases I didn't and resorted to turning off my vpn and looking at the thread, it was a mix of adverts pushing their product masquerading as comments, deleted or edited relics of the exodus, straith up wrong suggestions, and in general it didn't help me at all.

It was only two years, and it was basically half nornal computer science classes, and half working with engines, making a game with classmates and mentors from the industry throughout the year, and learning about rendering, AI behaviors (the videogame kind, not LLMs). The graphics part was about shaders, lighting, post-processing, global illumination, renderers and math, not modeling. It was mostly technical, but we had some game desing classes.

Having AI not bullshiting will require an entirely different set of algorithms than LLM, or ML in general. ML by design aproximates answers, and you don't use it for anything that's deterministic and has a correct answer. So, in that rwgard, we're basically at square 0.

You can keep on slapping a bunch of checks on top of random text prediction it gives you, but if you have a way of checking if something is really true for every case imaginable, then you can probably just use that to instead generate the reply, and it can't be something that's also ML/random.

I forgot to add that I had a Masters in Game Development and Computer Graphics, which definitely helped, but I still learned most of my gamedev skills by regularly attending gamejams and working on my own projects. I've also started working in gamedev for the past year, and I wouldn't say that it teaches you much, since you are missing out on 80% of actuall development and only crunch JIRA tickets and bugfixes, as a junior that is, without being exposed to the more important parts or other skills. Assuming you join a larger studio with game in progress, in an indie studio with team of 10 people, you'll probably have a lot more responsibilities and impact on other stages of the game's development.

This isn't true. If you can get by while working part-time, you still have at least 40 hours every two weeks to work on your game.

It's one of my biggest regrets, that after school I immediately jumped into full-time job, even though I realistically could live comfortably with 1/3 of the pay I was getting, since young+no familly+no car+shared living reduces your living costs by a very large margin. My best friend did that and has been working only 2 days per week since. I was trying to keep up with him, working on our game in my free time, but it's simply not feasible to build on top of 40 hours per week of regular job, and then do anything meaningful on your side projects. I barely struggled to get myself to do at least 20h of work per month on the project, missing deadlines, and it sucked.

He, on the other hand, kept our game project afloat and moving forward, with 60+ hours per month, while also writing and running a large LARP for 100 of players, directing his own theater group, and in general successfully working on a lot of projects, including several smaller games.

The best advice I can give, if you want to be a game developer, is to 1) not work in gamedev and 2) work part-time. The IT salary should net you a comfortable life even on part-time pay, assuming it's not gamedev. Smaller studios will have difficulties keeping afloat if they need to pay you, and in larger AAA studio you will be the same code-monkey crunching JIRA tickets as you would be in any IT job, but for a lot less money. And the design freedom you get when your livelyhood doesn't depend on your art's success, be it games or anything else, is totally worth it.

For example, this game has been developed solely in free time, without anyone getting paid for working on it. It's not AAA and the development takes a long time, but it definitely doesn't need to be a fulltime job.

I just ignore everything, avoid news, especially local or politics. I just don't care, don't plan to ever have children and just hope I'll get to live the rest of my life with my circle of friends playing games without having to deal with any kind of large crisis that would affect me.

Since I have slightly above avarage salary as someone working in IT, I'm counting on not being rich enough to be of interrest, while also not being poor enough (taking my lack of any expenses on family/car/etc into account) that if the living conditions worsen significatly, it will have already been a problem for more than half of a population way earlier and something will have to be done about it. And even if not, I can still comfortably get by even if prices of everything got 3-4 times as much as they are now, so IDGAF. It's a privilege, but I'm at a point where I don't really care what happens to others. For my part, I'm not bringing children into this hopeless mess, and while it's sad that a lot of innocent will take the fall, I also take solace in that a lot of the people who brought it on themselves will suffer for it.

What I mean by that is that it doesn't shove the AI summaries into your face, and they are only generated if you actually click on a different tab.

I just use Kagi, which seems to be pretty good at filtering bullshit by default, and have mabually downranked reddit and twitter, ot any other site I found and don't like. But it's been a long time since I used other search, so I can't compare it much since I'm used to it. Never really had any problems with not finding what I need.

I'm using Kagi, but as of right now I'm not sure if I can recommend it. The last year with it was amazing, but for the past few days I've been getting blocked searches from my VPN out of nowhere. That would be a dealbreaker for me, but I hope it was just a mistake and they will fix it. It's the first time it has happened in the year or so I've been using it.

Apparently, they are also adding a bunch of AI features, but I only noticed it when I was looking up the feature page, and I haven't noticed any of it in my feed before that - so I guess they don't push it on users and it's optional somewhere out of the way, so don't let that discourage you. (Though, it would've discouraged me, if I saw that before I started using it. But as of now it doesn't affect you unless you look for it, I guess)

Other than that, the search is awesome. But since I'm using it exclusively for like a year, I can't really compare it with other engines, it's possible that I'm just used to it.

I've never had issues with looking anything up. By downranking Reddit and using a search engine with a good indexer that downranks bullshit and generated websites, which mine is really good at, I haven't noticed much change from how it was before.

But I agree with the second part. That's something that never occured to me, and it makes sense. I was usually trying to answer questions I knew, and never had the urge to reply "just google it", so it doesn't change much for me, but it's a really good point I never realized.

I highly recommend reading Digital Minimalism, which deals with exactly what you are talking about. It's a great and inspiring read, even if you don't actually go through with it.

From what I remember, it mostly talks about how to approach any kind of technology as a tool, though a pretty simple process - honestly think about what your goal is (networking, getting information about new topics, keeping up to date on events...), and properly decide whether the technology is actually The Best way how to do it, while minimalizing any drawbacks.

Some examples I remember are:

• I have to use social networks to stay in touch with my friends and family. - Is that really the best way? Isn't it better to make sure to visit or call them regularly, so they can tell you what they have been up to, instead of you passively seeing it from soulless posts? If your goal is to have a meaningful connection with them, is chatting in a group-chat the best way and better than talking less often, but in person?
• I have to stay up-to-date on news. - Is endlessly scrolling through clickbait articles for hours the best use of your time? Wouldn't just subscribing to a physical newspaper/journal, that you get to read every day/week/month, accomplish the same thing while also saving you an immense amount of time?
• I need to have a smartphone so I can be reached at any time. - Do you need to have a smartphone for that? Isn't dumb phone way better, since you don't get distracted with other stuff? And are you sure that you have to reply immediately, and it won't wait until the evening? Will setting up regular 30 minutes per day, to check and answer your personal mails/IMs be sufficient, and if someone needs to reach you immediately, they can always call you.

Most of the arguments in the book were thought-provoking, and from what I've tried implementing, it has made my life a lot better. For example, switching my phone to a dumb phone (and carrying a powered off smarthphone that I can make a hotpost for, if I really need an app for something) made my away-from-computer life a lot better and peacful, and it was really easy to get used to that. Once you start considering anything you do on a computer from the pragmatic point of view, and ask yourself what your goal is, and if there isn't a better way - the answer usually is yes, there is.

You don't, that's the point.

Since it's a CAPTCHA (I haven't actually checked, I'm going by what you've written, I can't access reddit on my VPN), which I suppose they push to prevent bot scraping of their data so they can sell it themselves, the server checks whether you have completed the challenge it sends you with every request. I don't really know the details, but it should be hard to complete properly by a bot, since it watches a lot of things about how you interact with the browser and based on it decides if you are a bot or not. Which also makes it hard to fake the data, since faking it would be done with a bot, which they are pretty good at detecting.

If you don't send it, the server will refuse to talk to you. There is now way around it, other than paying for and using the Reddit API to access posts.

You can use a browser like Mullvad or LibreWolf, that should limit the number of things they can get about you, and using a VPN would go a long way. But VPNs have been recently banned from reddit (at least mine was), and it probably doesn't matter anyway - Reddit is tracking so much about you in background, that the few more data you send to them will not change much.

Such are the consequences of getting addicted to a large corporate product. Why shouldn't they do it? It's more data to sell, and what are the users going to do about it? Leave?

If I ever wanted to fight against my local regime, it would definitely not be through US and CIA, lol.

This puts to words something I was recently thinking about pretty well, especially the part about being an "advice seeker" and not really being able to solve stuff on your own, which is something I've always attributed to just being a field where you are driven to, especially in school, to have The Correct Solution, and that one always exists.

I mostly struggled with this when I tried getting into art, especially music or drawing. Suddenly, there's no algorothm or The Solution, and you have to figure out something based only on your creativity and judgement, and there's no-one who will tell you "this is the correct answer", which for someone being used to there mostly being one, was something I never managed to get over to this day, because it simply stresses me to the point of creative paralysis.

Thankfully, due to enshitiffication of most of the services I was following, which basically forced me to drop them due to invasive privacy rules, AI integration, or not working in privacy focused browsers or over a VPN, it's getting better. I'm kind of looking forward to OpenAI, Google and Meta finally killing most of the internet, so I can let go when 90% of content is AI generated, 60% of websites wont work without chrome, and the rest is just porn.

I find that kind of unlikely. If they wanted to frame someone just to have a killer, they wouldn't be talking about a "3D printed ghost gun", but just use a regular gun. I, for one, haven't known that it's possible to 3D print a pretty well working, and silenced, gun. And that might inspire someone - acquiring weapons is the harder part of any such murder, assuming you don't want to get caught, and the fact that you can get it without anyone knowing about it makes it way easier.

I admit I'm kinda disappointed. He pulled out almost perfect assassination that looked well thought out, managed to get away with only a few hickups in his plan as far as his face is considered, and then walks around with a murder weapon and a manifesto in his bag? Shame, really. All he needed was to lay low for a while, grow a beard and he'd probably be OK.