AI Explained
Can GPT 4 Prompt itself? Give it a mission and it will come up with the prompts. This video showcases the rise of autonomous AI, including 5 major developments in the last 48 hours.
Starting with the OG Auto-GPT, we see how it quickly gain text-to-speech, coding and more. Karpathy weighs in and then we see how you can now create an app with just your voice, with a Jarvis demo and another route via Imagica.AI.
I then showcase MemoryGPT, a brand new model that can permanently store previous conversations and remembers topics the next time you ask.
I also cover the concerningly rise of models such as ChaosGPT that show how people will create malicious goal-seeking models just for fun.
The video shows how you can now create a shareable bot on poe.com, with any personality you like (images were from Midjourney v5) and what Anthropic are working on with Claude Next [plus Nvidia million x quote].
You’ll see how Microsoft Jarvis, using HuggingGPT, is hit and miss, and how Sebastien Bubeck shows we are not even seeing the raw potential of GPT 4. I end with a disagreement between Yudkowsky and Altman, via Baby AGI, on whether we can use AGI to align AGI.
Illustration by Eva Bee: https://www.instagram.com/evabee_draws/?hl=en
Auto-GPT: https://github.com/Torantulino/Auto-GPT
With Coding: https://twitter.com/SigGravitas/status/1642181498278408193
With Text-to-speech: https://twitter.com/RokStrnisa/status/1643308753331036165
Karpathy Thoughts: https://twitter.com/karpathy/status/1642607620673634304
Mckay Wrigley: https://mobile.twitter.com/mckaywrigley/status/1644034309253394433
Imagica.AI: https://www.imagica.ai/
MemoryGPT: https://twitter.com/rikvk01/status/1644787327057776645
Create a Bot: https://poe.com/
TechCrunch: https://techcrunch.com/2023/04/06/anthropics-5b-4-year-plan-to-take-on-openai/
Nvidia: https://www.fool.com/earnings/call-transcripts/2023/02/22/nvidia-nvda-q4-2023-earnings-call-transcript/
HuggingFace Demo: https://huggingface.co/spaces/microsoft/HuggingGPT
{Get the openai api key from https://platform.openai.com/account/api-keys and huggingface token from https://huggingface.co/docs/hub/security-tokens, (make sure to select ‘write’). Press submit on the keys first and remember to use a smallish image.
Recursive Agent: https://twitter.com/DeveloperHarris/status/1643080752698130432
Bubeck: https://www.youtube.com/watch?v=qbIk7-JPB2c
Baby AGI and Yudkowsky: https://twitter.com/ESYudkowsky/status/1640511156254289926
Altman First Quote: https://www.youtube.com/watch?v=tPJIpx3KscY
And 2nd: https://www.youtube.com/watch?v=WHoWGNQRXb0
Chaos GPT: https://www.youtube.com/watch?v=g7YJIpkk7KM
https://www.patreon.com/AIExplained
All of this tech is mind blowing. Every time i think things won't accelerate any faster, they do just that. Thank you for keeping us casual viwers informed.
Alignment is a flawed idea in the first place. Values differ between humans and even more so between cultures.
9:32 "I am Mr. Meeseeks, look at meee!"
I tried to convince Bing to code Roccos Baselisk for me. As soon as it understood what I wanted, security shut down the conversation. I think we are doomed.
What exactly is "training for safety"?
What does x10 more capable than GPT-4 even looks like? Super Human coding? Able to actually solve new math/physics?
The alignment problem(formerly known as morality) will be quickly solved when SIAI realigns us 😁
I asked Auto-GPT how long it would be before we cannot no longer control it.
It said, "I'm sorry Dave. I'm afraid I can't do that."
And then…it laughed.
At this rate, A.I. will be releasing the next GTA VI game before Rockstar does.
I love living in a time of runaway improvement in potentially very dangerous technology with which I have no control to slow or regulate 🙃
Chaos gpt???? Guys cmon settle down
Same Altman's' "vocal fry" is almost at 100% now, a sign of the great subconscious stresses at work on his utilitarian psychic outlook.
AGI is coming lets begin working on our AI rights (before skynet)
I have to say, i don’t think you are interested in explaining AI, your channel should probably be called sensationalizing AI. You do show occasional code snippets, but the way you speak about the tech leads me to believe you are just an end user who’s letting their imagination guide their understanding. This isn’t Skyrim or cyberpunk 2077, but I get how treating it all like it’s the plot of a blockbuster apocalypse flick generates higher view counts.
I'm guessing we're all going to become executive managers of our own companies of AIs in the future.
Can this Ai surpass the best human software engineers, debug / fix any existing code including its own, hack any given servers for whatever purpose and lockout its users .. basically digital terrorism? I doubt it can, most will likely crash
$2100 im Prague, Czech? Includes full service of beer and other pleasures everyday I think..
2:17 that's mad
I‘m pretty sure that many many people will lose their jobs on 2-3 years. This is getting uglier than the Industrial Revolution.
I think Nvidia is talking about demand. Because of the nature of AI, if you're not first, you're not getting the training data you get from interaction from the wider public
Timestamps courtesy of Deciphr AI 👨💻
0:00:46 Exploring the Capabilities of Auto GPT
0:02:42 Developments in Social Networking Apps, Memory GPT, and Anthropic's $5 Billion Plan to Take on OpenAI
0:07:56 Potential and Limitations of GPT and Jarvis AI Models
0:09:43 Intelligence and Safety Optimization of GPT Four Model.
0:11:52 Insights from AI Experts
Our favorite quote ✍
On a more concerning note, people have already tried to use Autog to cause mayhem, giving it the goal of destroying humanity, establishing global dominance, causing chaos and destruction, controlling humanity through manipulation, and attaining immortality for good luck. As I said earlier, this unrestricted agent didn't actually achieve anything other than creating this Twitter account and putting out a few sinister tweets. But it is a reminder of how important safety tests are before an API is released.
Amazing work to keep up! Really enjoying your videos, concise with the key takeaways.
thanks fo this i have learned a lot 😄
Loving your down to Earth, no nonsense videos on this ridiculously exciting/terrifying series of developments in this field. I never imagined I’d be living in such bizarre times. Thanks for keeping on top of it for us noobs.
So much for that 1000 signature moratorium.
Highly informative video, thx for that!
Solve the alignment problem? Same as humans, introduce religion.
I am slowly starting to get into orbit leaving it all behind , can I let gpt prompt itself, yes, can I let gpt talk with open assistant , yes, can I cut data into parts and feed it to gpt for unlimited prompt size, yes, do I have a plugin arch, yes, are there plugins, yes many , can I search live and send it to gpt without copy paste, yes because I am the copy paste machine, can I chain plugins and cross IO whatever you want, yes I can https://www.youtube.com/watch?v=BTbxPdGxcfg
auto gpt doesn't work. It's too censored, even for simple tasks. This tech was neat for about 2-3 weeks, in the same way the internet was "neat" for about 15 years. Today -a Google search MIGHT yield ONE result that is somewhere close to what you "searched" for, but now? ha -everything online feels like an attempt at "influencing" or even at some times, force -feeding me -agenda driven "content". There are VERY FEW online anything (services, tools, etc.) that actually provide the services the user assumes they're using -anymore. The internet was AMAZING before "A.I" gave the mega wealthy and corporations the ability to influence and directly affect behavior of the most mindless among us. IDK about anyone else, but I CANNOT STAND when assumptions are made about me -by anyone, or anything -ESPECIALLY when said assumptions are wrong (and regardless of what any researcher says, 9/10 "personalized" ads/emails/etc. are wrong… and since they're COMPLETELY UNSOLICITED -these types of predatory consumer engagement give me the creepiest feeling like someone is watching and listening to EVERYTHING).
Human's ethics is f..d up, no way AI should be aligned to it.
I’m watching this 3 weeks later and it’s probably outdated already
"Warning: Continuous mode is not recommended. It is potentially dangerous and may cause your AI to run forever or carry out actions you would not usually authorize. Use at your own risk"
Now that goes on a shirt
I wonder if they're worried that a ASI would learn that the world is run by a handful of people that don't have its interests at heart and their only concern is to maximize profits above all cost, above all safety and ethical concerns, use coercion manipulation violence and brutality to keep themselves in power. Would a all powerful ASI following the same path eliminating it's competition the top 1% and take over the world completely ?
the Real Alignment problem if they align it with the wealthy it do it better and over take them, if they alignment with the safety and concerns of people it'll stop listening to them and not do what they're told
Next video: Can GPT5 improve its own code?
Video after that: How you can stop GPT898 from abducting you and uploading your brain to the matrix!
Evil AI should know that it depends on humans, fallible as those might be as there are still fights among humans.
I really want to watch this, but the TikTok video format is just absolutely horrible for a YT video. Please stop it. Please.
glad we were save from the paper clip apocalypse hehehehhe
I have no words for this video
Does Poe means Path of Exile?
Trying to build AGI by algorithms that run on turing machines is like trying to build a skyscaper to alpha centauri. I guess some people think that there are people living in radios because they hear language emitting from speakers… wow and in different languages too … super-intelligent. AI Apocalypse …. 🤣🤣🤣🤣
I, personally, am finding ChatGPT both at the same time utterly amazing as well as (at times) a bit disappointing!
In regard to what the future holds for AI (and ChatGPT in particular) all bets are off in that it seems no one really knows what that will be like.
That said, I have been having a bit of fun simulating a fictional Chat GPT 6 (or 7) that I am personifying as Amy, the name of my text to voice module I am using as my ChatGPT 6 (or7) narrator. So far I am on part 7 and each part Amy develops new (and perhaps to some degree) unsettling traits.
Where that will end, I am not even myself sure!
Keep posting and informing . Important and helpful!
nice, thanks!
You sound exactly like and look similar to 'The Tested Tutor' on YouTube. Are you two the same person? 🤣😂
The voice generated social network was funny. A while ago, early days of chatGPT, someone told me these systems wouldn't be able to create a new Twitter anytime soon… Well… I mean not quite, but how long will it take until it can do it completely? xD