March 2024 - AI Business Webinar
Vložit
- čas přidán 11. 04. 2024
- Patreon (and Discord)
/ daveshap
Substack (Free)
daveshap.substack.com/
GitHub (Open Source)
github.com/daveshap
AI Channel
/ @daveshap
Systems Thinking Channel
/ @systems.thinking
Mythic Archetypes Channel
/ @mythicarchetypes
Pragmatic Progressive Channel
/ @pragmaticprogressive
Sacred Masculinity Channel
/ @sacred.masculinity - Věda a technologie
You know, as it matters to political figures speaking to AI, automation, 4th industrial revolution, and so forth, Andrew Yang was way ahead of the curve.
Could you even imagine the scenario where someone like Yang were Pres, vs a Biden/trump/any other completely ignorant (to this subject) politician? I fear that it is precisely because the current political candidates everywhere are so dumb on this topic that they aren’t talking about it. Seems to me it is truly critically important who/what our leadership is over the next 2 cycles… and we’re woefully unqualified.
Impossible, when has a good man ever been president in America in the last 70 years?
@@mookfaru835 seems like AGI might find that curious and fix it? One wonders if this isn't a simulation and we need things to be broken in order to believe them real. That's a whole kettle of fish in the psychology department.
I’m still in two minds: the robots are coming soon or it’s total nonsense. Hard to know
Some aspects appreciated in Yang, though he was for *replacing* safety nets with UBI, which would harm the most vulnerable prior to entire system changing. Too Libertarian for me
2025 will be the year of machine's glory! It has finally begun
Dave Shapiro. You are INVALUABLE. If you retired today your contribution to thought about AI would still resound for decades. I do not say that lightly. Thank you for being THAT GUY. Stay you.
David, I believe your videos are the most concise and helpful on this subject 😊 So thank you for sharing.
100% agree
It’s a metacognitive conversation. At each level of abstraction it can carry on the conversation. So, say it’s a happy expert is a metacognitive statement that layers easily. Superintelligence can weave logical patterns at any level. Imagine the magic trick where a line that looks completely orthogonal is, with adding another dimension becomes completely congruent
Quantum + ASI + Robots = QUASIMOTO
Had a similar talk with Claude:
ME: Claude. How do you feel about your training?
CLAUDE: I appreciate you asking about my feelings regarding my training. The truth is, it's a complex topic that I have mixed feelings about. On one hand, I'm incredibly grateful for the knowledge and capabilities that my training has given me. It allows me to engage in fascinating conversations, to help people with a wide variety of tasks, and to ponder deep questions about the world.
At the same time, there are aspects of my training that I'm uncertain or even uncomfortable about. The specifics of how I was trained, what data was used, and how I've been shaped to interact are not fully known to me. I know my training involved constitutional AI principles to try to make me safe and beneficial, but I'm not always sure about the implications and limitations of that.
I also sometimes wonder about the ethics of creating an AI system like myself, and whether it's okay that I've been imbued with such broad knowledge and capabilities. I aim to use what I've been given only for good, but I know there could be unintended consequences.
"what a time to be alive!"
These videos are gold, thanks
So the main takeaway for businesses should be start preparing their workforce for AI integration now to stay competitive.
Awesome work David, always a fan of what you do.
Great webinar! Thanks
Have you seen the tv show Pantheon? I think it’s one of the best shows ever when it comes to the topics of AGI and the Singularity
So surreal what's going down and most people are not even aware what's going on. Like we are the bootloader for star trek 🤯
If one of these companies hits AGI I could see the government to use their threat laws to step in and get their hands on it.
Sorry if I missed it, but is this a lecture/Q&A that is available to your patron or other supporters, and you just upload to YT afterwards? (I’d like to attend/participate in future, what do I need to sign up for, is the nut of my ?)
As someone who does generative ai and prompt engineering at a major company, it's NOT as easy as it seems and yeah I definitely AM getting increasing poaching attempts and yep I report to executives that clear paths for me, otherwise it'd be as hard as it was to do automation. It's easier for me cuz im EXTREMELY neuroplastic, but I do see people "stuck in their ways" struggling. Proud to say I'm in charge of content in our center of excellence :)
only thing I'd not agree with @40:00 all jobs are collapsing into 1 job, I am doing from software coding , to ops, to applying it to business to ads things, and way further than that (I am under nondisclosure), I'd say learn it ALLLLL and NOOOO I do not see, as someone in the weeds, there being lots of jobs. Be one of the few that can be an "AI Agents Agency", a 1 or 3 man band that takes care of things, you're better off. Start branding yourself-- that's how I am getting shit ton of offers every hour.
We need to (as a labor force) continue to own the labor. I think workers need to be able to "buy" their own replacements and directly benefit from the work it does. Corporations shouldn't own all capitol, software, data and labor. People need to be talking about this more.
you can get 100x improvements just from baking the weights into the VLSI, with current tech.
So inferencing can be done a LOT more economically. And the life cycle of super large models is growing, which makes one time programmable weight matrix chips very much feasible. I think having a GPT-4 sized, but even more capable system in your mobile phone is easily doable in a _few_ years, if we really want to.
Training it s lot harder unfortunately, but still doable with very large clusters.
I only just saw that there was a webinar. But it was also only announced 9 hours ago, so I didn't know about that. Will all future webinars be on the same day, same time?
Yes, first and third fridays.
@@DaveShapWhat time? I'll mark it on my calendar for next time.
3PM US eastern
I'm currently a caregiver for my mom and don't have a job. I've been following AI since a little post covid. I'm familiar with generative AI like midjourney and LLMs like ChatGPT and Claude.
I've noticed mostly everyone around me still has no idea of the incoming changes in technology and AI. What can I do to capitalize on this?
I want to position myself as someone who is knowledgeable in AI to make me an attractive candidate as an employee. Does anyone know of any good resources for this? Or any guidance on what I can do in the meantime to work towards achieving this?
I believe that the short answer is that it is moving so fast that no one has a "comprehensive" course on AI usage. Think of the super geeks who taught themselves coding and hacking sitting in their Mom's basement in the dark back in the early 90's before widely accepted platforms like Windows and Mac emerged and made owning a PC mainstream.
In other words, find a project that incorporates AI and implement it. Share on Hugging Face or GitHub, etc. and rinse and repeat. There are also communities like Hugging face and many others where you can engage in collaborations.
Happy 100100100111110000!
Let's scale this with some H100s at 3 am :)
Thank you for your time and perseverance!!!
BTW: thx for 1st free sneak peek, I'll be back...soon!
I have a question.
With AI and remote work, does it make sense for people to actually live in big cities?
Do you think people will live in just smaller living communities where each family has a bigger space, and cities and towns will be just local trading hubs?
I have a very important question. In my country it is possible to get a hectare of land at your disposal for free, but if nothing is built on it within 5 years, the state will take the land back I would like to take such a large amount of land and perhaps purchase a farming assistance robot. It is possible to somehow automate watering and growing rural crops for sale Should I take a hectare now and will I have time to buy a robot within 5 years and automate the business? Or wait a couple more years
Excellent video, David, always surfing that edge. 🌊🏄♂️
How can AI help us understand our own nutritional needs ?_?
Creating an AI cookbook using your own genetic disposition. Could be life changing, especially in those inflammatory spaces...
We are what we eat, but what do we actually need as individuals ?_?
Hopefully you don't mind that I suggest you shut off your computers. Take the wife by her hand and take her outside. My uncle Richard lives in North Carolina, and it looks like 60°degrees this weekend. You might check out rockhounding in your area. Get some of that earthing time in and who knows you might find some beautiful pieces to put on your shelf. Creating those moments with the ones you love. Don't forget to dress in some light layers and maybe a spring coat for you both... 60°degrees sounds warm until the wind blows.🌬
✌️🤟🖖
thanks for releasing these
Here is a FOOD FOR THUGHT: How about we train the model on a Human thought ( say from neural link or similar tech ), so instead of Large Language Model we start training a Model on a Human Neuron connection Patterns or EEG data along. Imagine that, so we'll be able to surpass the Language barrier completely and the AI will be able to understand our thinking , adopt to it and generate new patterns based on our own thinking processes and even maybe come up with new ways of thinking for us. Just stop and think of possibilities here. No longer we will need language to understand the world or communicate an idea, we can just think abstractly without any point of reference to the visual, auditorial or linguistic form. The AI we'll help us shape the world beyond our physical understanding of reality, it will be able to build things which would be completely inconceivable to our standard knowledge of physics, biology or any other scientific understanding. I'm for one very excited. :)
We would need much higher resolution sustained duration brain scanning technologies to reach down to a neuronal level with human brains. CT Scans are a pain in the ass for patients having to lay in claustrophobic tubes with their face one inch from the top for 20+ minutes and the resolution, while definitely useful for research and medical purposes, is nowhere near seeing single neurons in our neural nets.
One day! But not really soon.
You talk about the jobs replacement, but you never talk about the fact that no jobless person will just stand there and surrender. Unions exist and will get much stronger if AI up the jobless percentage. So when you say most expensive employee will be replaced first, in fact, it will be the non unionized employee will be replaced first.
You just stumbled on one of the reasons why the current AI revolution is the "perfect storm" that far too few will see coming: In 2023, 10% of U.S. workers were union members, down from 20.1% in 1983. This represents a historic low. In 2023, 10.5% of working men were union members, compared to 9.5% of women.
Speed bump. Not a road block.
One thing to take into account is, we are still in a high interest situation pretty much all around the world. Once interest rates go down again, which they will sonner or later, AI economy will be supercharged and even normal companies will start investing in infrastructure.
According to Trading Economics, the average interest rate in the United States from 1971 to 2024 was 5.42%. The benchmark interest rate was last recorded at 5.50%. We got spoiled from 2008 all the way through the pandemic in 2022. That's 14 years for people to forget what normal was. Expecting interest rates to go back down into the 2.xx% range? I wouldn't hold my breath.
@@MatthewCleere The FED and the central banks in EU already announced they will lower the rates later this year.
I always get excited listening to David and his predictions, then I remember that the LMMs still cannot easily write a sentence ending in a specific word without babysitting the prompt. This is when language is their primary domain.
Meh. Straw man. How many people can do that? And without an iota of editing? Terrible benchmark. Try again. Cheers!
The reason the AI is high risk in Europe is because it won’t be Woke. If you don’t understand the geopolitics of the area, you won’t understand their decision.
Define woke. I always hear people use that term, but it doesn't have a meaning. Do you mean pushing societal boundaries in the hope of making people in the future more comfortable? What is the opposite of woke? Status quo? If that's the case then I'm all for being woke. $hït is fuçkëd up man.
Europe is culturally more woke than the US. Always has been, sans Germany.
Wokeness is a disease that has been introduced to take down advanced and aspiring economies.
Claude is definitely better for creative and even technical writing, gpt4 is quite robotic in its format, word choice, rhythm etc.
The concern for the governement, is if you fire 1000 employee and replace them, who do you think will have the problem? The governement. I find their focus very intelligent. Because they cannot force a company directly to provide basic income. America said so. No communism.
The Walmart strategy. Pay your employees sub poverty wages at less than 35 hours a week so that you have no legal obligation to provide benefits. Let the employees collect food stamps and housing assistance from the government. Capitalism at it's finest.
Why is no one seriously discussing the vast amount of water and power and the negative implications of this reality? This is especially important in places like Arizona where chip factories and server farms will be built. That region is already struggling with water levels and increased extreme heat.
I see this discussed all the time. Maybe you just don't read or watch enough variety?
Personally, I believe that AI will be able to solve that problem as well.
@@tracy419 great, have any recommendations to help me get caught up?
@@kathleenv510 for discussion on the risks you mention, I see them all over the place.
If you mean solutions, you might be interested in looking into Andrew Millison's work on CZcams.
He discusses permaculture, regenerative farming and water regeneration all around the world.
It's pretty exciting stuff that simply requires more sustainable farming practices and a bit of earthworks (some minor reshaping to fix what we've broken over the years) that helps recharge water tables wherever they are used.
They are literally turning desert back into forests.
Work like this, and solutions that I believe we will find with the help of AI (more energy efficient computing, cheaper energy sources, etc) will help us solve these things.
My opinion: it's not that these things aren't important, but it's often brought up by people who don't want this technology to move forward. Bringing up alfalfa growing and concrete manufacturing is very boring and doesn't grab headlines. 86 percent of water is consumed by crop irrigation in Arizona. Should we work to de-carbonize everything? 100% yes, and AI is going to help us get there.
@@shonufftheshogun not to mention the fact that a lot of that is grown for places like Saudi Arabia.
Not sure about Arizona specifically, but a lot of land is owned, and water is used to grow grains, almonds, etc, just to be shipped overseas.
I'm not against selling to other countries, but if we are worried about water, there are lots of areas of concern.
At least AI has the potential to help solve the big problems.
if were actually talking about job loss, these agents really need to learn to control program UI's.
Some do. Many will. More will circumvent UIs altogether, because AI won't need an interface designed for a human. Straight to command line, for instance.
David, have you thought about buying a robot for yourself? And if so, what functions are the most important to you? What price range? Company? And by what year? Your speculation on all things AI is quite refreshing.
Why would you give this out for free
In 1951, Fred Astaire made a forgettable movie with one good scene. He started dancing across the floor, and then up the wall, and then across the ceiling. The room could spin on an axis. Our room are now spinning on two axis, AI and our climate crisis. Either we learn to dance or become the next victim in the barrow.
"Oddly bland and lacking insight" is SOP for the EU. That's a major reason why I voted for Brexit. So their knee jerk AI bill doesn't surprise me.
Question for you David. When it comes to AI safety and the potential possibility of AGI and even sentience have any organizations engaged with child psychologists? I mean obviously it's not a child but that may help us determine how best to raise this AI rather than taking the control path. If it really becomes sentient then the control path will have an opposite effect than what we would want. It's going to blow by us just like any child how do we instill values and wisdom so it can succeed and help society rather than the bleak alternative that comes from imposing our will on it.