General Discussion
Related: Editorials & Other Articles, Issue Forums, Alliance Forums, Region ForumsGPT4: "Last night we got new superpowers"
GPT-4 launched. It was explosive. Here are some of the best examples of it
Yesterday was our Everything Everywhere all at Once moment for AI. Google, OpenAI, MidJourney and others launched new AI products.
But, it was OpenAIs launch of GPT-4 that blew my mind.
When Apple launched the iPod showing us we could put 1000 songs in our pocket or the iPhone showing us the future of mobile, it felt big.
OpenAI gave us GPT-4 creating a website from a napkin drawing. It felt transformative.
Work is changed forever for everyone when you can go from idea to finished website/app with no steps needed in between. Product is not longer the blocker for great ideas.
-snip-
GPT-4 is even more transformative for creating media. It can handle 25,000 words! It can easily work with you to write books, compose songs or research papers.
With 25,000 words AI will be much more powerful. For example, you can ask it to consume all of David Ogilvys books on copywriting and use what its learned to rewrite landing pages, sales pages, your home page.
No person can keep 25,000 words in their head, and effectively reason and manipulate with the information. The impact here will be incredible.
It will transform education. I can give GPT-4 whole sets of documents and ask it to provide a step-by-step guide to teach me whats in it woah. Khan Academy just launched an AI tutor for their courses.
-snip-
Its multi-modal so can accept visuals as an input. Its image recognition is incredible. It can take any image input and describe it perfectly. Being able to have images as an input along with text will be huge.
Presentations will also be disrupted. Tome.app showed an example of their app creating an entire presentation from a Wikipedia article using GPT-4.
Coding continues to be transformed. We saw an example of GPT-4 being given code with bugs, consuming documentation on the code, and then perfectly correcting the errors. Why learn when GPT-4 can do it for you.
-snip-
https://www.linkedin.com/feed/update/urn:li:activity: 7041763466428194817
Working URL: https://tinyurl.com/anm4vfew
Here is the launch video:
brush
(53,764 posts)highplainsdem
(48,959 posts)Renew Deal
(81,852 posts)What skills will be important in a future that includes generative AI? I think writing, critical thinking, and communication skills will always be important. The future will also require skills to work with AI, identify inaccuracies, and find ways to stand out in a world dominated by a few AI systems.
highplainsdem
(48,959 posts)heard about, and all the damage it can do.
Potential damage that OpenAI is well aware of, but they just keep saying policymakers should regulate it to lessen the damage, while giving them no chance to catch up.
Those motherfuckers are in an AI arms race, and they don't care who gets hurt, or how badly.
brush
(53,764 posts)There's potential for thousands of jobs to be lost as decades of institutional knowledge can be sent packing.
Am I reading the OP correctly that website building/design will be able to be done quickly by AI? That may no seem a bad thing to some as site building. can be tedious, but what happens when a site crashes? Will there be no web master to come to the rescue?
And this concern just touched briefly one field. There's academics of course, other areas of tech, scientific research, and on and on.
highplainsdem
(48,959 posts)suggested we'll need Universal Basic Income for all the unemployed people, he's delusional about getting to a UBI.
Saw a profile of him done a few years ago where he suggested a family of 4 could survive on as little as $3500 a year. Because we'll have fusion so power will be free, and free power will mean food will be almost free, and kids can learn anything they need to know from their smartphones, won't need college.
He did admit the $3500 wouldn't include housing costs. As I recall, he was at a Silicon Valley mansion at the time.
And that's the "genius" running OpenAI.
See this thread, and reply 9.
https://www.democraticunderground.com/100217685127
brush
(53,764 posts)I knew things would go sideways as Musk is rumored to be extreme right in his views (many of his actions/policies indicate it), and Altman is allegedly way left of that, clashes would invariably occur.
And neither is an Einstein, but it's pretty obvious both think they are. Altman expresses tremendous visionary exuberance but nevertheless hedges his bets by stocking a remote getaway with all the provisions he'll need in case AI gets out of control and humans lose control of it.
He doesn't inspire confidence (hears hoping his views have evolved since then), and Muck certainly does not.
Renew Deal
(81,852 posts)There's a lot of info about what it can do here: https://openai.com/research/gpt-4
You are correct that there is an AI arms race. And I doubt anyone in the US will regulate it any time soon because the US has gotten ahead of China in the public arms race. China is reported to be ahead of the US on AI overall. I doubt anyone will attempt serious regulation in the US in the current climate. There's a chance something will be attempted in Europe, but that is a big risk for Europe because they might get left behind if they try. Societies need to find a balance. Get the most out of AI without the avoidable impacts. AI decision making for credit and critical services is an avoidable potentially harmful impact. The ending of certain job categories is an unavoidable impact, as it has been throughout human history.
highplainsdem
(48,959 posts)Renew Deal
(81,852 posts)Or at least some time experimenting. It didn't reach 1 million users in 5 days and 100 million in 2 months by accident. There are some real benefits such as improving text, providing or supplementing ideas, answering questions, entertainment, and more. It's also very efficient. There are also risks such as inaccuracies and non-answers (for certain topics). But there are risks with any tool and with all people.
LudwigPastorius
(9,128 posts)GPT-4: "What do you mean we?"
Buns_of_Fire
(17,173 posts)- from Colossus: The Forbin Project, a 1970 film
Or, if you're into something a bit more cheeky:
User to GPT-4: "Is there a God?"
GPT-4: "There is now."