Fxiz

@MostlyHarmless when I was majoring in AI, people around me were building models for early Alzheimer detection, hyper-resolution in medical imaging, traffic optimization in smartcities, plant disease detection and many more. Many were just college projects and not on par to be deployed, many were. But you know what was the difference there? People were seeing AI as independent tool. Not a property owned by corporate giants. That's why they were able to see it's use case. Can you see it as well?

@lproven does it mean if I build an app on .net and push the build on my phone, it will not work? Nah. That can't be true. Ohh wait a minute. I use a ~7 year old nokia phone. Only God can send an update on it now.

@VeroniqueB99@mastodon.social can it be ethical and environmentally stable? Maybe Yes. Does it enable great things that people might want? Yes Is it made by well meaning people for good reason? "Well meaning people?" What kind of dish is that? Now, can you analyse why the answers are yes? No. Because "no" is what I read in a post so mine is also no. I won't think about it myself.

@admiralwonderboat@mastodon.social and you people ask for "free will" and freedom in everything? Imagine being scared of automobile. Why, because they unemployed cart pullers, who ironically now transformed their job to some other sector. People are not even ready to understand what the tool means, and they act as if they have known it all. Using water, data centers, artists, stolen data, yeah.. memorized these keywords by heart to be true "know it all". No need to ever know about the rest of it

@pwloftus @sundogplanets I would agree that infra and setup will be getting a huge chunk out of the total earning. But that is as shareholder. A system builds on methodology before equipments See, my opinion, we have a ton lot of generalised models, but those are for hype. To build uses. And now that there is hype, and requirment, you would be seeing a pattern that now companies are not building LLMs, they are integrating micro AI is general system workflow. As for truth, we can ask them(?)

@pwloftus @sundogplanets let's take it this way. You were a very fit person. You walked 10km every day with no strain. I build a product and gave it to you. Now you don't need to walk 10km every day. Few years later, you are no longer fit enough to go 10km. What I gave you? A car. It's not just cars, pick up anything from past to present. Everything was build to make human life easy, in process making humans less capable than day before. Who all should we blame it on?

@nowayeast @rickyx @nixCraft what is secure then? If you call phishing hacking then if a person gets your bank account details and withdraw cash, will it be "bank system hacked" or "bank security is dead"? Bank got hacked or you got hacked?

@iam_jfnklstrm @dozymoe @mrundkvist yes indeed nice to exchange thoughts. I wish a good day to you too

@iam_jfnklstrm @dozymoe @mrundkvist yeah. Everyone has their own definition. The only reason AI is not able to reach the level of intelligence you are thinking is the peak, is because the kind of AI we use are ment for it. LLM models/chatgpt is more about answering and answering correctly. It does not support a natural flow of conversation. But that it can be build, their is agentic AI. Such AI are not available because they are use case specific, and requires even better level of AI development

@iam_jfnklstrm @dozymoe @mrundkvist let me explain myself Martin. Humans take information in, they learn, they store that information in some form in brain. When a new situation comes, they search for that information and make up inference. In principle, AI is same. It trains on data, it learn, and it performs inference based on that learning. And obviously, with inappropriate learning, or no learning output will be wrong. Same wrong output you will get from children with no learning.

@theeclecticdyslexic @GeePawHill just for ethics part, it's totally on human. I don't think it's very wise to blame a tool for it. That is in itself offloading accountibiity. Or like putting cart before horse. Let's build accountibiity for how this is build and used, rather than miss guiding it towards "block AI, problem solved". Something else will replace it.

@itsfoss@mastodon.social true. It especially breaks when you starts dreaming about making it "beautiful". I use arch, it works as I want, but as I install a DE, it done for. Maybe I become too ambitious at times with it.

@MachineLordZero it's not about a day. When rich kids celebrate their birthdays lavishly, even kids from humble family sees it and wishes to have a day to be celebrated in their names. It is as simple as that.

@theeclecticdyslexic @GeePawHill ohh you mean to say cooking pot? It's a device I can copy, and surely if one try hard enough it can use to do to human existence deed. So what should we do, ban pots? But then people will start calling it "monopoly", "fascist" and what not. If a thing is public, it's problem, if it's centralised, it's a problem. Surely you see which one is worse right? Remember, the good old penicillin. Glad it was not banned. If you say conversation is over, it is I guess.

@hootalex while I understand the point in the article and I would agree to save wildlife as much as possible, I can't comprehend meaning of the text in the post. "Tech Bros?" The guy is building a house. A house, not a tech firm. He just so happen to be also a tech entrepreneur. Don't we have a term in English for such collective discrimination based on specific group of people?

@torgo People here has reached such high level of understanding that they have known each every thing. They are predicting things that no one with best of the knowledge is able to. They are validating things. And deciding what is good and what is bad. Even after being in tech field I am unable to say things for sure but out here people are giving verdict left and right To people against AI and tech people: no one takes you seriously only because you people speak with no information. Know first

@raymaccarthy @mcjk @gael my friend there is a concept of investers. A lot of things exists in this world, but not many knows about them. Someone invested in it to make it that popular, and that's why many knows them now. Investment bought the fast and wide development of the tools. How it went down the road and how people sees it now does not invalidate the whole history.

@dendari @raymaccarthy @mcjk @gael well... You live in/build a house to prevent yourself from natures harm, not to bath in hail storm while sleeping in a cold winter night. No one is Saint. Imagine going to office 9-5 and at the end of the month they send in your bank account a "You did a good job. Everyone is happy with you."

@theeclecticdyslexic @GeePawHill argument? You mean to say word play? Sorry my friend. My vocabulary is not that vast. Knife is a tool. Knife has a lots of uses. Some can use it awesomely, some cut themselves. And in some bad cases it is used for worst, but knife is not a murderer. If you can't even understand that, then all you are best at is argument.

@ra6bit @zackwhittaker I am sure not many cares about it. It has the keyword "AI" with a negetive narrative, that's all what matters. That's how new outlets write news headings, that's people now a days gain information and call themselves all knowing.

@Microplastics101 @passwordsarehard4 @brainwagon those situations are called accidents/bad machinery my friend. LLMs veer right? And those are intentional? Would love see those claims proven in court. People on internet go around throwing words all over the place. Go ahead and challenge it. If successful, you will be able to shutdown those firms. Your heart desire would also be fulfilled. Win-win situation for you if you ask me.

@torgo bunch of 'gods' here on social media deciding what is good and what is wrong Sitting at home, having a sip of coffee, seeing some company's ceo on tv, who might have worked day and night to reach there, these 'gods' go: "hmm, I don't like it. Why are they ceo? Why are they earning more than me? I don't like it" or "why is this guy using machines in his company? It's 'his' company but I say he employee humans, so he should! I don't like it" That's your comment area, in summary

@sundogplanets @pwloftus I appreciate the thought mate. But that is just another article. I can quote dozen here which says they are making money. Then how will we decide which one is telling the truth? Let's just speculate then. Ethically. Let's say I am running a million dollar company. I personally wouldn't want to loose money. And even if I agree to bleed money for now, I might want that bleeding to pay back at least 3 fold. Makes sense or am I being too generous?

@mrundkvist that is the problem. While one is using statistics (I think not), other is using chemical reactions to perform those same operation for identifying those matching patterns. Human brain is working with chemical reactions and electrical signals, just for argument sake, even more unreliable way to do it. But it's not right. We are over simplifying that AI uses math, but it is essentially just trying to stimulate same kind of operation as that happen in brain.

@Microplastics101 @passwordsarehard4 @brainwagon no car is a fine example. If a person drunk drives and harm people, car won't be going to jail or held accountable. And if it does, that's counter productive and misplaced responsibility. Or should I instead call it, "a machine that is designed to move fast in a certain way, designed to go in whatever direction the driver wants is different. Cars have been shown to perform accidents intended movements." That's narrative.

@iris_meredith@mastodon.social then what is the problem with this setup? If it's just ambitious start-up with no substance, then they will just burn their money, status quo breaks, they, you, both same now. If it's a start up that actually works, they are no nepobabies, they are inventors. I don't think there is any problem with some being rich, and intelligent, or is there?

@iam_jfnklstrm @dozymoe @mrundkvist aren't you romanticising the term intelligence a lot? You are saying humans can reason while ai form opinions based on slop from internet. The Cambridge dictionary definition, you pulled that up from internet as well. And you, human, made reason based on it. How is it any different from ai's opinion from internet slop? All condenced down to just word play to set up a narrative.

@theeclecticdyslexic @GeePawHill multiple drugs/ medicine came from unethical testing. Now common medical practices were once experimented unethicaly. The phones used to post this runs on semiconductor material acquired from mining which disrupts nature and carbon footprint. Electricity comes from hydroelectric plants, that disrupt aquatic like. Many more. But those are fine now as those are not in trend and you are also using them right? Online anti-ai sentiment deserves those fancy vocabulary

@LRRRonEarth from "ai is slop" to "kill people who are building it," truly goes on to shows the kind of mentality you people have. Reminds me of a story from a certain book..

@iam_jfnklstrm @dozymoe @mrundkvist fair enough. You are making your stance based on current state of AI and how it is been used. That's fair. I am just saying that maybe, corporate is doing things wrong, AI is not the best at the moment, and everything. But it has potential. Are we agreeing on that? And potential can only be realised if people actually try to do it right instead of just disregarding it completely. We know the mistake with fake watch, now we can add fallback check for it

@vanzick @nowayeast @rickyx technicaly I beg to differ from the use of term hacking here. I would use different vocabulary, "vulnerability of user." If the argument was about the user system, then yeah it can be categorised as hacking, but not other one. Otherwise I can store bank details on open cloud and call it bank got hacked. As for old bank system, a sign or fingerprint was more than enough for transaction. Heard many stories about people impersonating those. You wouldn't want that back.

@iris_meredith@mastodon.social what kind of services are we talking about? Clubs, pubs or casinos? Because other kind of services are designed to be used by general public and get money out of them. If it's a service for general public and it's bad, we don't use it, they don't earn, they loose business, go broke. If it's a good service, we use it and no one need to care if it is from nepobabies. And if it's a service for riches, they play with it themselves, no concern for general public as it's their money

@pwloftus @sundogplanets That's what I am curious about as well. My estimate of real world model with ~400B parameters and ~10²⁵ FLOPs would be: Training- Ideal case: ~500MWh Real world: ~2-5GWh Usage(10 text promt of ~200 tokens and 10 images)- MOE: 10KJ to 100KJ Standardized: 0.0028-0.028kWh Even if current GPT model has ~1 trillion parameters, very likely MOE, it won't be using all parameter at once. "2.9kWh"? For 10 response and 10 images? Nah! Corporate wants to earn, not that dumb.