How not to be replaced by AGI Keep up with AI Job Market – AI Tutorials – Timestamp: 00:00 Introduction 00:44 Research and Development 07:59: Technical Features 17:32 Responsible AI 25:36 Economics 30:50 Science & Medicine 32:21 Education 32:36 Regulation 35:18 Public Opinion 36:19 Summary š¤ Follow me on Twitter š Order My Website – Links from today’s video: Welcome to my channel where I bring you the latest discoveries in AI . From deep learning to robotics, I cover it all. My videos offer valuable insights and perspectives that will expand your knowledge and understanding of this rapidly evolving field. Be sure to subscribe and stay up to date with my latest videos. Was there something I missed? (For Business Inquiries) contact@theaigrid.com #LLM #Largelanguagemodel…
Crap
You lost me when you said the government should get involved. NO sane person should ever want the government involved. Of course there will be a power imbalance. There ALWAYS is and should be with new technology. That's infinitely better than having the government slow everything down while trying to stop such an imbalance.
Why are you talking so doggone FAST!!!
"AI wont impact certain industries." Let me know which ones so I can change careers to rhat because it seems to me that there isnt ANY industry that wont be impacted. I mean with robots becoming more functional and implementing AI, and eventually cost effective, I feel like that was the biggest statement of the video.
Not all political opinions are worth representing.
My personal prediction about the expected minimum level of automation for jobs not yet automated as of April 2024, which are currently occupied by humans (future novel professions are not included in this prediction):
2024: 0.01+%
2025: 0.5+%
2026: 3+% (possibly the first AGI)
2027 – 2028: 9+% (at least 1/7 of all office jobs are affected, first applications for real full autonomus programming becomes public available, first general multipurpose robot becomes ready for mass production)
2028 – 2031: 20+% (at least 1/3 of all office jobs, during this interval the Cambrian explosion of robotics also begins, possibly the first ASI)
2031 – 2035: 50+%
2035 – 2038: 70+%
2038 – 2042: 80+%
=> I estimate that at some point during the given intervals, at least the specified degree of automation will be reached for jobs that are occupied by humans as of April 2024.
Get the government involved?? Lol.
Oh, all LLM are definitely biased.
Regarding ChatGPT politics… well, reality has a liberal bias. That's all there is to it.
I find it strange that Suno isn't on the music gen report. Sure, v3 is somewhat new, but Suno Chirp has been around for almost 10 months…
For the copyright issue… look at the prompt. The prompt is literally commanding Midjourney with the unique name of a copyrighted character, the name of the copyrighted work that character's from, and TWO instances of "make sure this is a screenshot of a scene directly from the movie". Did you know that if you ask a human artist "paint Starry Night by Van Gough, precisely that painting," they can also reproduce existing works? Only difference is that the diffusion model is incapable of refusing an order. For typical uses where you don't demand the AI violate copyright, it won't. I don't see this as a tech issue so much as a user abuse issue.
LLMs are smart enough to tell if a request would violate copyright; so maybe we should be putting an LLM in front of every art-generating AI model and instruct it to refuse such prompts.
In terms of political bias… let's all remember that there's a difference between "unbiased" and "balanced across all opinions". False balances are worse than someone/an AI not agreeing with a particular user. I'm not saying these AI models aren't leaning in one direction or another, I'm just saying the goal shouldn't be "make sure the AI espouses all views equally". Maybe there's a reason that alignment training designed to make the AI more friendly and less harmful results in one political leaning or another, besides "all AI companies are liberal cucks trying to silence dissent" š¤·āā
Rip its been fun
Will the future continue to be a battleground or are we going to turn it into a playground? If the latter, read Amaranthine
I suggest reading Amaranthine: How to create a regenerative civilization using artificial intelligence
We really need to rethink what it means to be human not just the future of AI. If we release AI into a predatory corporate society…
"Explicit content is something you don't really want"
Speak for yourself. I agree with you deep fakes should be illegal, but the "kids can't watch porn so nobody should be allowed to" is part of the problem in society.
Business dominant and academic heavily linked to business following. This is a fatal combination, we only have to look what business is doing to our democracy and the disaster which the pharmaceutical industry has become to realise that this is a worst case senario. The US leading in what appears to be a race with China. The actual position of China will not be as clear as made out in this video by nature of the Country. Of course this all suggests that AI is under humanities control which ultimately it will not.
Sometimes being unbiased means taking a side that is closer to certain political position, being in the middle doesnt mean you're unbiased, it just means you are in the middle, sometimes reality is biased.
Please do not say yadaa, yadaa, yadda. instead etc, Certainly! Here are some synonyms for "yada yada yada":
1. Blah blah blah
2. Etcetera
3. So on and so forth
4. And so on
5. And the like
6. Et cetera
7. And such
8. And more of the same
9. And the rest
10. And similar expressions
ā¤
The government is working on it come on!š theyāre already there. These other companies are just to get you used to AI
Why-TF are your opening captions so blurry??? I swear I started this video three times, and then began to wonder whether or not I was having a stroke.
Mind officially blown! AI keeps pushing the boundaries of what's possible. Can't wait to see what's next!
What constitutes political bias? Are the claims of bias themselves biased? Claims of fact needed be verifiable. Opinions are just opinions and should not carry a great deal of weight. The criteria for judging policy should be that they do no harm. All actions can be judged as either good or evil. Even an AI should be able to classify any given statement.
Microsoft and Google are US based so if you count in industries US is ofc leadingā¦ point is, is it actually US leading? Do these major corporations even have something to do with their original country anymore? The future Iām looking at is corporations, itās a sci fi reality at this point, major power imbalance between private lead companies and governments..
I'm more worried about the actual politicians than their deep fakes and more worried about the white hats than the black hats
You trained to talk slower – very good! I had to do the same.
great video, thank you
i think i can hear some kind of fan or something in the room, your mic is picking up a lot of the sorrounding sound, or maybe im crazy
9:50 Bro literally said "Humans are going to be better than AI in pretty much everything by next year"
BRO STOP SAYING YOU KNOW, PLEASE.
It is funny how young males talk about AI companies like they were college teams in the playoffs.
STOP SAYING "You Know!!!!!!!" Every bloody sentence.
Canāt wait for agi to arrive! #Iloveagi
Waitā¦9:52 āHumans are going to be smarter in AI.ā??
doom is nearby
wow this is it… April 17 2024 marks the date when AI is smarter than humans in some areas
Talking points covered in the video (via GEMINI):
1. AI Index Report: This report tracks and visualizes data related to AI, aiming to provide unbiased information for policymakers, researchers, and the general public. The video covers 10 sections of the report.
2. Industry Dominates Frontier AI Research: In 2023, industry produced more notable machine learning models (51) than academia (15), with a trend of increasing industry dominance. The speaker expresses concerns about the potential negative consequences of private companies leading the way in AI research.
3. Open-Source AI is Increasing: There is a significant increase in open-source foundation models (149 in 2023, double the amount in 2022). This trend might be impacted by future regulations on open-source systems.
4. Training Cost of AI Models: The video estimates the training cost of various models. For instance, GPT-4ās training cost is estimated at $78 million, which is less than expected. The speaker finds this interesting because closed-source models tend to be state-of-the-art.
5. Leading Countries in AI Research: The United States is the leader in AI research, followed by China, France, and Germany. The speaker wonders if this trend will continue, considering factors like potential regulations and investments in other countries.
6. AI Patents Are Increasing: There has been a significant increase (62%) in the number of AI patents granted worldwide, indicating a rise in AI investment.
7. Multimodal AI Advancements: Recent advancements have led to the development of strong multimodal models that can handle various data types, like text, image, audio (e.g., Googleās Gemini 1.5 Pro). The performance of these models on multimodal benchmarks has significantly improved, reaching a human baseline of 9.04% on the MMLU benchmark.
8. New Challenging Benchmarks for AI: As AI models reach performance saturation on established benchmarks, new challenging benchmarks are emerging to assess their capabilities. The video mentions the SWE bench for coding, Heim for image generation, MMu for general reasoning, Mocker for moral reasoning, and the Agent Bench for agent-based behavior.
9. Human Evaluation of AI: The video discusses the LMS chatbot arena for LLMs, which uses human evaluation (e.g., ELO rating) to assess AI models. GPT-4 Turbo is currently leading in this area. This type of evaluation is valuable because it considers user experience, not just performance on specific tests.
10. Impact of AI on Robotics: LLMs are being integrated with robotics, resulting in more flexible robotic systems like Palm-E and RT2. These models can also scrape content from the internet and potentially be used to create misinformation campaigns.
11. Responsible AI: The video mentions the importance of responsible AI development and discusses potential biases in AI systems. It emphasizes the need for unbiased AI models to avoid influencing usersā political views.
12. Impact of AI on the Economy: The video discusses the economic impact of AI, including potential job displacement and the need for policies to address these concerns. Investment in generative AI is rapidly growing, with some concerns about a potential bubble.
13. Public Perception of AI: The video presents a survey that shows that while 60% of people have heard of ChatGPT, there is still a significant portion of the population that is unaware of AI advancements.
Thanks very much for your channel mate! Appreciate your contribution.
Can AI finally turn your 37 minutes of talking into a couple minutes of to-the-point content?
Repent of your sins or suffer the consequences. Lord Jesus died in our places personally to take the death punishment that sin deserves and then resurrected by the power of God. Believe this and sincerely repent of your sins each time you sin and you will have eternal life and nothing to fear. Fail to repent and you will end up in the Lake of Fire.