jacksilver
@jacksilver@lemmy.world
- Comment on Google Kneecaps Loads Of Very Big Websites After SEO Change 6 days ago:
If there were more search systems/engines there would be a wider variety of ways search results are optimized. Meaning SEO would have a greater level of diminishing returns. Having a single player creates a single point of weakness in search.
- Comment on Odours have a complex topography, and it’s been mapped by AI 3 weeks ago:
AI has been co-opted by all the GenAI people. The number of times I’ve heard things like “AI is the next big thing for X business” and they’re only talking about GenAI is way too high.
- Comment on Adobe Firefly used thousands of Midjourney images in training its 'ethical AI' model 3 weeks ago:
The phrasing was poor, but what Adobe did is effectively the money laundering version of copyright theft. Just churn it a couple of times and then the original creators are now “AI models” and not the peoe whose work was used to create the “AI models”
- Comment on Movie industry demands US law requiring ISPs to block piracy websites 4 weeks ago:
Another non-Canadian who found the show by happenstance and think it’s great! Also watched the animated seasons when they came out (although not quite the same).
- Comment on AI Has Lost Its Magic 5 weeks ago:
I think it’s amazing and terrible at the same time. It clearly produces some amazing looking things, but I’ve never been able to get it to create what I want.
- Comment on Valve: Windows 11 market share on Steam drops to 41.61% 5 weeks ago:
Even if their sample size is large, you’re assuming that people who are willing to respond to the survey are evenly distributed across different OS. I wouldn’t be surprised if their is a slant to this data just because it’s based on an opt in survey.
- Comment on What are the best indie games you've ever played? 1 month ago:
I picked up Balatro because of this thread and I agree, it’s a great game and something a little fresh, but slay the spire is still probably the best.
I agree about the “more like a game” element. Baltoro feels more like playing cards than playing a video game. I think it’s cause I’m using the same odds/play styles as when I play real life card games.
- Comment on What are the best indie games you've ever played? 1 month ago:
They’ve definitely added things to it over the year or so it’s been out. Not sure if enough to make it worth it for you. It’s also possible I’m just bad at it, as I haven’t beaten it (although only 12hrs on it).
- Comment on What are the best indie games you've ever played? 1 month ago:
Wildfrost is a really good one. It’s got a lot of different play styles and has a pretty big learning curve.
Also on android there is Pirate Outlaws. It feels like a slay the spire clone, but has a good amount of content and does enough different to be worth mentioning.
Not exactly the same cause it’s not a deck builder, but has a similar feel is dicey dungeons (both steam and android). It’s a lot simpler and luck is a larger factor, but it’s got a decent gameplay loop and being able to play on Android helped scratch that itch on the road.
- Comment on Leaked SpaceX documents show company forbids employees to sell stock if it deems they've misbehaved 1 month ago:
They’ve literally made reusable boosters and have multiple operations to deploy satellites and go to the space station (not to mention starlink ) . While I’m not a Musk fan, those achievements are irrefutable. SpaceX may or may not be making money, but it’s a far cry from a ponzi scheme. It’s why so many are trying to copy their technological achievements.
- Comment on Google says the AI-focused Pixel 8 can’t run its latest smartphone AI models 1 month ago:
Doesn’t that require all of those things to be developed manually though. It’s not like the LLM can just access your logs. I guess the thought is the LLM can better generalize commands, but seems like a lot of the AI there would actually be hand developed.
- Comment on Google says the AI-focused Pixel 8 can’t run its latest smartphone AI models 1 month ago:
That explains why Google wants it, but what do phone owners get? What is being offered of value. I keep hearing all this talk about this wave of AI on phones, but don’t see what it’s providing.
- Comment on AI Prompt Engineering Is Dead 2 months ago:
Actually most (I think all, but not 99% positive) machine learning models are incapable of doing straight arithmetic. Due to the way they are built ML models, including deep learning models, can only learn relationships in a limited input space.
This is most apparent when you test LLMs on different arithmetic operations:
- For addition, it does okay up until you get to millions or billions
- Multiplication I think breaks at the 100/1000 level
- exponents almost break immediately
- Give it decimal values and it also breaks relatively quickly for any operation.
This has to do with the fact that LLMs are effectively multiple layers of linear functions, so higher order operations break down faster.
- Comment on Air Canada must honor refund policy invented by airline’s chatbot 2 months ago:
That actually already exists and it’s terrifyingly stupid.
- Comment on Hogwarts Legacy has officially cleared Zelda as 2023's best-selling game worldwide 2 months ago:
Haha I appreciate the comment and the ability to call them out even though you like them.
I just wish I felt the same. The longer they’ve been out the more I realize that we probably won’t get a more traditional zelda ever again. I think the thing I liked about zelda up to BOTW was that the world itself was a puzzle. Figuring out how to navigate and open up new areas was part of the fun and challenge to me. Not to mention dungeons being larger and more intricate puzzles than anything you come across in BOTW and TOTK.
- Comment on Hogwarts Legacy has officially cleared Zelda as 2023's best-selling game worldwide 2 months ago:
I’m really confused by all of the story comments in this thread. It’s fair to criticise HL’s story, but at least there is a story and characters. What story does TOTK even have? What characters have more than a line or two? While Zelda has never been big on complex narratives, at least previous entries (before BOTW and TOTK) could develop a story since they could have a linear progression. A couple of flashback scenes really doesn’t tell a great or compelling narrative and really disconnects the gameplay from the events going on.
- Comment on Plex for books? 2 months ago:
I’ve used Ubooquity on and off for this. It has some nice features, but isn’t open source.
- Comment on Hogwarts Legacy has officially cleared Zelda as 2023's best-selling game worldwide 2 months ago:
I’m curious, what open world games do you rate as a 9 or 10? I’m not saying Hogwarts did anything revolutionary, but it did most things pretty solidly. It’s been a while since Ive played an open world game that does a good job on making the world actually feel alive.
- Comment on Hogwarts Legacy has officially cleared Zelda as 2023's best-selling game worldwide 2 months ago:
I’m kinda curious in what way Zelda (assuming TOTK) has more depth. Combat wise HP has stealth, an attack typing system, comboing, special moves, and more if I recall correctly. TOTK does have a variety of weapons and you can craft weapons, but it generally boils down to just whacking away at things. You could also mention the ability to make vehicles/automaton, but the time to build things (until you find ultra hand?) mixed with limited resources made that more of a pain/chore than fun.
I could go into other mechanics, but ultimately I think TOTK would be rated worse if it wasn’t for the Zelda branding carrying it.
- Comment on Where are the good political songs? 2 months ago:
I didn’t see this mentioned so wanted to add it. Most people think of it as pro USA, but the lyrics paint a very different message.
For example:
Got in a little hometown jam So they put a rifle in my hand Sent me off to a foreign land To go and kill the yellow man
- Comment on Article suggests that 1 million ML specialists will be needed in 2027. What do you think of that? 2 months ago:
It depends on what you’re calling AI. The LLM hype may die down, but Ml/AI in general has been continuing to grow and expand for well over a decade. It’s just unlikely that all the things being prophesied right now will come to fruition.
- Comment on Adobe Photoshop's AI tools put women politicians in bikini bottoms and their male colleagues in suits 3 months ago:
This isn’t about nonconsenual images, it’s about bias in AI models. They used the extend image feature in both images and because the models think women=sexy it produces them in bikini bottoms and men=business it puts them in suits.
This is going to be an ongoing issue in how generative AI assumes things based on the prompt/input image - www.bloomberg.com/…/2023-generative-ai-bias/
- Comment on AI Companies Lose $190 Billion After Dismal Financial Reports 3 months ago:
LLMs don’t really fall under AGI, they’re still static statistical models. Some RL algorithms might be on the track of AGI, but I’m not sure about that.
- Comment on CEOs say generative AI will result in job cuts in 2024 3 months ago:
And how does it know how to process the request? Someone still needs to program or enable it to do things, otherwise you’d just be talking to a wall. So maybe you end up with less prompts, but the end state is still going to be terrible service.
- Comment on "Did you realize that we live in a reality where SciHub is illegal, and OpenAI is not?" 3 months ago:
I am familiar with how LLMs work and are trained. I’ve been using transformers for years.
The core question I’d ask is, if the copyrighted material isn’t essential to the model, why don’t they just train the models without that data? If it is core to the model, then can you really say they aren’t derivative of that content?
I’m not saying that the models don’t do something more, just that the more is built upon copyrighted material. In any other commercial situation, you’d have to license/get approval for the underlying content if you were packaging it up. When sampling music, for example, the output will differ greatly from the original song, but because you are building off someone else’s work you must compensate them.
Its why content laundering is a great term. The models intermix so much data that it’s hard to know if the content originated from copyrighted materials. Just like how money laundering is trying to make it difficult to determine if the money comes from illicit sources.
- Comment on "Did you realize that we live in a reality where SciHub is illegal, and OpenAI is not?" 3 months ago:
Actually content laundering is the best term I’ve heard to describe the process. Just like money laundering, you no longer know the source and know it’s technically legal to use and distribute.
I mean, if the copyrighted content wasn’t so critical, they would train models without it. Their essentially derivative works, but no one wants to acknowledge it because it would either require changing our copyright laws or make this potentially lucrative and important work illegal.
- Comment on The New York Times sues OpenAI and Microsoft for copyright infringement 4 months ago:
Except no one is claiming that LLMs are the problem, they’re claiming GPT, or more specifically GPTs training data, is the problem. Transformer models still have a lot of potential, but the question the NYT is asking is “can you just takes anyone else’s work to train them”.
- Comment on Best games that can be completed in under ten hours? 4 months ago:
Its a really fun game and one where youll actually enjoy 100% it.
- Comment on The first minds to be controlled by generative AI will live inside video games 4 months ago:
But being able to talk about anything and having the character actually do something based on the conversation are completely different things. Yeah you can convince a random npc to “join your quest”, but unless that was programmed into the game the dialouge and the actions of the npc will contradict each other (making a worse interaction).
NPC dialouge is purposefully limited to align with what the game is programmed to do, we’re still a ways away from really being able to leverage the advances in LLMs in video games (at least based on what I’ve seen).
- Comment on Why do AI image generators have a stroke when they try to generate text? 4 months ago:
To add to this, the way the AI is trained is that you pass in images with descriptions (for the most part). Since most descriptions focus on the main concepts, it generally won’t have the actual text included in the descriptions. Without the being included in the descriptions, the AI will have a hard time learning the meaning of the squiggles in the images. In addition those squiggles can represent a lot of different things, so even if it grows to “understand” letters, it’s really hard to “understand” their meaning; thus leading to a lot of weird words/text.