this post was submitted on 10 Feb 2024
505 points (100.0% liked)
Technology
67338 readers
4931 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related news or articles.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
For perspective, global annual GDP is $105 trillion, which means Altman is asking the world to invest 6.7% or so of the entire world's economic output for one year in his company.
Holy fucking bonkers when you put it that way. Like holy fuck.
Are they that close to something amazing, or is Altman going true Dr Evil megalomaniac?
It's both. Basically, if he's right that this is among the most important tech in world history and deserves $7 trillion in research, it'll make OpenAI the du jour monopolist over said most important tech in world history if he gets it. Outright dystopian.
Generative AI is not an important development. General AI would be. This is just a party trick.
Yeah, AI's been a whole field for many decades, machine learning and deep learning is a whole field that existed and was doing tons of interesting work across a variety of domains before transformers came out. Fully agreed that LLM's are a party trick, but companies use party tricks to gin up interest and money for their real work all the time. None of this changes the fact that AI is important technology.
It's also open source too. With faster chips and standard training sets, it will be trivial for anyone to train a basic AI to recognize fire hydrants or whatever. Just like with computers, the "revolution" will not be one giant company but every business using their own.
Your fridge only needs to recognize food items. It doesn't need any more intelligent software. I want a food recognition AI and cameras in the fridge and cabinets so I know what food I need to buy. This doesn't even need live cameras (for power consumption, privacy, and storage). It can just take a picture and analyze it when the door opens, then delete the picture.
Sorry what's all open source?
Llms. There are very competitive, fully open source models.
Openai is not.
They’re bullshit. Any breakthrough in General AI world be news everywhere forever.
Aside from the fact that OpenAI and Microsoft leak like a sieve, the scramble for investors would publicize the development, and also the fact that Altman is a braggadocious idiot who need everyone to think he’s smart or he’ll collapse in on himself. The world would know, there isn’t an argument to the contrary that holds any water given the scale of the breakthrough general AI would be and the actors involved here.
The breakthrough was the ability to do like 3rd grade math. Being able to reason is a big step to researchers but it’s still very much in the research phase.
They're as close as theranos was, just need the money and "two more years".
theranos never changed the world, openai already has
Eh, not really in a positive way yet. I use it for quick code snippets but it’s not all that important compared to the fact that misuse of generative A.I. is making so much information on the Internet completely worthless. I don’t think the trade-offs have been worth anything we’ve seen yet.
I’m bullish about a lot of generative A.I. use cases long term but they haven’t accomplished much positive yet and there’s still major, major scaling issues that may not be solved soon. It could easily be like self driving cars where progress stagnates. A lot of times with software, the first 80% is easy and the last 20% takes a decade longer than people expect.
That change has mostly br a bunch of empty promises. There is no AI yet, there is machine learning, which is just an ingredient. As impressive as they are, they're mostly useless
Machine learning is AI
I don't know if creating and tying yourself to a bubble that's inevitably going to burst is exactly a great way to change the world.
It's a great way to make a fat pile of cash though.
Ye and he wants it for a fucking chatbot.
They are that close to cocaine.
And the world's 5 biggest companies their total market cap is a bit over 8 trillion dollars, which is also mind boggling, that so much capital is concentrated in the hands of so few people or shall we say mega corporations.
Came here to say this. Does she know how much money there is in the world? He is asking for basically 1/20th of all the money in the world. Even if that was possible it would be dangerous for one company to hold that much.
Money can be created by banks via loans, you can just make it up.
But then you end up with inflation. So in the process of creating money you've reduced the value of the money. And banks working in cooperation with government create money. Private companies outside of the banking system can't create their own money anymore. If any sizable portion of this is taken out as a loan. It creates a systemic risk for the world economy.
Banks most certainly make up money today via loans. This happens all the time.
But this isn't a bank. And making money doesn't come without consequences. You're not thinking about second and third order effects. This would basically be quantitative easing on a grand scale but for just one company. It would literally destroy the economy if the investment failed. And they aren't the only player in the industry. The level of systemic risk is too large. And if it didn't fail, it would basically be handing the world economy over to one player.