it so over…we just don’t see it yet.
4 Likes
I am totallly just rocking as Company4 at this point
4 Likes
The reality is starting to sink in:
OpenAI CEO Sam Altman has gone on record saying that an energy breakthrough, like nuclear fusion, is needed to make advanced AI viable.
This viewpoint is supported by some very serious studies. One from the University of Massachusetts Amherst looked at the computation and energy costs associated with improving an image recognition AI performance to over 95% accuracy. They found that training such a model would cost $100 billion and produce as much carbon emissions as New York City does in a month. Bearing in mind that this is for an AI that still gets it catastrophically wrong 5% of the time.
from: AI Is Hitting A Hard Ceiling It Can’t Pass
If this new AI model would cost $100B to train, can you imagine the cost of a monthly subscription? You can fill a room of Flame artists at full rate, and it would be a lot cheaper.
The illusion is complete.
Unfortunately when this bubble bursts, the blast radius will be bigger than some of the recent bubbles close to the tech sector.
3 Likes
It’s already happened. He is an investor.
He addresses this in the above article. To make fusion competitive (price-wise) with current energy production is a tall order that a lot of people smarter than me don’t think we’ll see in our lifetimes. From what I can tell as AI stands currently, the model requires a lot of food and it’s getting to be too obese to be sustainable. There are so many reasons for a potential plateau (running out of quality training data after scraping the entire internet, more and more energy usage, more and more expensive) but simultaneously so many reasons ($$$) for people to keep evangelizing.
Great article. Thanks.
OK. Now Autodesk is in the game themselves, on the 3D front:
And the responses have been spot on…
3 Likes
Cool – maybe the 2D department will get the memo at some point.
I’m a lawyer, so I’ll talk from the legal point of view.
AI has proven to be immensely helpful in legal research and document analysis. It can quickly sift through vast amounts of data, identifying relevant cases, statutes, and regulations much faster than a human could.
While AI can assist in processing information, it’s the human touch that is essential in legal interpretation and ethical considerations.
Dr. Nick Oberheiden, the founder of Oberheiden P.C., highlighted in this article the importance of using AI to improve our efficiency and analytical depth while assuring that the human element remains at the forefront of our efforts. So I agree with him. AI is revolutionizing the legal industry by streamlining processes and increasing efficiency, but it’s important to remember that it’s a tool to extend human capabilities, not replace them.
2 Likes
There are so many great uses for it. We, of course, are hung up on what it means to us. I can easily imagine it concatenating data from hundreds of traffic cams and adjusting traffic lights to optimize flow.
2 Likes
Except when it invents cases, as it’s been repeatedly caught doing. Have they fixed that yet?
3 Likes
My brother works in the state department and we were talking about this the other day. A lot of the best use cases are not very sexy to the general public and really come down to having created a beautiful algorithm for dealing with big data sets. They don’t remove the need for human intervention, but get the ball rolling in the right direction for people to then double and triple check before revising and iterating and implementing. It seems like just a better algorithm for this being pushed as something that is going to fundamentally alter our day to day (when in actuality it might just be invisible, behind the scenes chugging and plugging data like a really wonderful excel spreadsheet on steroids). Cool.
The generative stuff (in comparison to the analytical) IS sexy to folks though. But to me that’s the most troubling aspect by far. The idea of never being able to trust something you see on the internet ever again is disturbing and a real bummer in terms of digital evolution as we move from the Information Age into what could likely be called the Misinformation Age (which it’s hard to deny we haven’t been moving towards for a while anyways). To be able to not trust our eyes is a fundamental unmooring of the experiential data we rely on in order not to go completely crazy and become ungrounded from reality. It’s not pleasant to think about and it’s exhausting to think of a world where you’re forced to constantly second guess every image you see.
2 Likes
Also add many documented cases of bias and racial profiling when AI is used to analyze information that is is used to make life changing decision, from mortgages to parole.
It’s very good at the common use cases, but does not favor minorities.
Thus human supervision is always needed. Not just on the decisions but also the data quality.
Not that humans are free from bias either. But automated mass bias that has no phone number to call is more problematic.
2 Likes
Apparently trucks carrying trees aren’t in training sets either…
A video of a Waymo car swerving like a drunken sailor in downtown Phoenix. Don’t quit your driving job yet 
https://www.instagram.com/p/C67BuGmyANs/
Goes to show that we’re all just guinea pigs sharing the road with that type of immature tech and no supervisor anywhere nearby. In most other industries that would be unacceptable. We’re roasting Boeing for much lesser offenses.
1 Like
and so it begins…
3 Likes
im gonna throw this really fascinating article onto the pile. I find it very difficult to argue the majority of the points made by Doug Shapiro here.
funny to see @ALan’s post above… basically proving the point of this article!
2 Likes
Sometimes AI is unnecessary.
There are people in the world that can stand in front of a camera and just invent all kinds of things in real time.
(And a small part of me openly wishes that AI would take their job…)
Thanks for sharing that… Watched the video, dude is 100% on point and tracks with what I’ve been thinking about for awhile. Next 5-10 years of entertainment is going to change drastically. Hollywood as we know it now will be gone.
1 Like
I don’t know. According to this article, in 2023, the AI industry spent $50B to generate only $3B in revenue. Unless that ratio drastically changes, the situation gives off massive mental masturbation vibes.
Thanks for sharing DD.
4 Likes