Quantcast
Viewing all articles
Browse latest Browse all 2602

Personal Investments • Re: Retiring in 6 months (off TSLA options); Plan Review; 08/24 Update - NVDA All-In


Yep. We have to come to terms with a few things

First the fact that direct to consumer AI has been a complete disaster. Gemini is a laughing stock. Every company that decided to do AI search results, chatGPT clones, etc spent billions in an ultra competitive market in order to then lose money on every customer because the models are so expensive to train and run...

Second that lprogress on more powerful models has stalled out. Hallucinations might not be solvable. Chat GPT 5 might take more data then currently exists...

I can't tell you what the future holds for AI, but it's entirely possible that it'll go the way of the metaverae, a money pit everyone will want to forget soon. Maybe it goes the other way, but the negativity is just starting to get priced in after all that optimism.
The effects of tech are often overestimated in the short term and underestimated in the long term. AI is much more than just self-driving, and chat gpt, and these are early days.

Computers can now see and hear. They can almost do once unthinkable things like... drive a car. We have a lot of jobs associated with tasks like driving cars, trucks, and trains, or reading and paying invoices, or operating cash registers and such.

I just read about Bosch developing AI's that listen to machinery to identify a malfunction before it happens (it was an aside in an article about a company making progress on the "cocktail party problem". One could imagine that someday, instead of security cameras going unmonitored, or having a guard try to periodically scan dozens of video feeds, security cameras will be able to flag an incident in progress and/or follow the suspect across cameras to tell you exactly where they are now, if they're still in camera range.

Costs will inevitably come down. Right now, it's early days, and researchers are still figuring out what is possible. ChatGPT just happened to work way better than anyone would've predicted for this early point. Enough so, that major companies (including non-tech companies) have FOMO that if they aren't working on this tech, they might end up disrupted by competitors.

It's anyone's guess, though, what this means long-term for any given company. Netscape and SUN microsystems were early leaders, but that didn't necessarily make them good long-term investments
I agree with your final paragraph strongly. I think your response is level headed and optimistic and i appreciate that. You may be more right than I am. I do want to dispute a few things since I work in the field:

First, AI is a BS term that makes machine learning sound new when in fact it's been a massive and expensive research focus for decades. We're not in the early days. Self driving cars, transcription algorithms, OCR, these have been around for a while now and the progress has been incrental and expensive. Computers can listen to you talk for a while now, but Siri didn't trigger trillion dollar arms race, so why are people pointing to transcription as a partial reason for the AI boom?

Second, I wouldn't put any stock in what any company is saying about what they're doing with AI. It's a buzzword and people know mentioning it will make their stock go up even if it's all smoke. Think about how every company accepted Bitcoin for like 6 months in 2017 because the optics of accepting Bitcoin were exciting to investors, then quietly dropped it when no one ever used it. What Bosch is doing might be real, it might just be a prototype that never sees the light of day. Either way ML algorithms like that aren't new at all. The current boom came about over generative AI models, but Bosch wouldn't be using generative amAI in such a system, it's probably a pretty bog standard regression model which have existed in production for a while.

Third. Costs will not "inevitably" come down. Remember when costs were inevitably going to come down on nuclear fusion? It's been 20 years away for 50 years now, and it's 20 years away now! The barriers to better AI are legitimate and not easily overcome. Chips really can't get any smaller. Microsoft gave OpenAI billions on the promise it would develop a cheap model. It failed and is burning cash at a higher rate than almost any startup ever.

Again I appreciate your optimism and futurism. I have strong hopes for the future as well and I do think it's cool when my computer writes code for me. My point isn't to say that AI is bad or worthless, but just that it has the potential to be something society basically skips over like the Metaverse. If that happens, OP may see his leveraged position go to zero. Maybe all the points I'm making end up being irrelevant and AI will usher in a new age. I don't know. But i do know my diversified total market portfolio is pretty ready for either scenario.

Statistics: Posted by cvoege — Sat Sep 07, 2024 8:40 pm



Viewing all articles
Browse latest Browse all 2602

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>