Does it seem like things are not going all that well with AI in general. From Googles image AI and now their generative AI flub

https://finance.yahoo.com/news/googles-generative-ai-fails-will-slowly-erode-our-trust-in-google-202314852.html

I didn't even ask what was going on but overheard my wife talking with a coworker about employees trying to use AI, and apparently it's obvious if not comical

Maybe this whole race is still a few years out. I get that you gotta have the chips first, like NVIDIA but their trajectory the last 6 months can only be compared to meme stocks.

Anyone else starting to take some cash off the table in their AI portfolio? For all I know it blasts off as soon as I submit this post but I'm not as short term enthusiastic as I was, software in particular

It probably depends on the industry. In general, I think current AI tools are great for idea generation but should not be relied upon for anything factual. I haven't seen any current tools that I think are close to being able to meaningfully replace the work of white collar positions, but each iteration of the tools seem to improve exponentially. I will be interested to see what entrepreneurs can build on top of these LLM and other AI models.

themissinglink said:

---

It probably depends on the industry. In general, I think current AI tools are great for idea generation but should not be relied upon for anything factual. I haven't seen any current tools that I think are close to being able to meaningfully replace the work of white collar positions, but each iteration of the tools seem to improve exponentially. I will be interested to see what entrepreneurs can build on top of these LLM and other AI models.

---

How would you relate that to the current price of AI stocks that aren't chips required up front

Lots of software side AI stocks are way up but not there yet. At least the chip makers are selling a crapload of product

I've seen lawyer friends use it. They can take a bunch of client files and get a summary of key people, plot, options, etc. They used to have to pay a paralegal to sort thru it all.

I can see it being useful in contract negotiations, supply chain strategies, etc. All that is prob few years out.

Right now I think fake news and viral clips are what we may see most of first. So whoever can develop AI recognition flags or become the AI snopes, they will do good.

If they are just now getting the infrastructure in place to support the large language models needed in various sectors, you shouldn't expect to see anything good for probably 3 years at soonest is my guess, and even that may be wildly optimistic. The payoff with AI is several years out for most industries if not longer.

I am a skeptic that AI as it's being talked about will deliver what some people are imagining, but we probably won't know until 2028 or 2029 if it was a bust for the early adopters.

Now….how all that timeline plays into investor fatigue, that I don't have a great idea on. But just knowing how the software life cycle works and how this is a newer frontier, it won't come without growing pains and some losers who overestimated it and some that underestimated it.

I am newer to the investing game but what all the people with more experience and wisdom keep telling me is you ride the winners until they aren't winners. And right now there's no doubt that there has been no better place to have your money than in the AI / semiconductor sectors.

Four Seasons Landscaping said:

---

OP sounds like a very new headline focused approach to AI.

The people that actually know how to use AI aren't really impacted by the "flubs" he mentions.

---

Agreed. AI is so much more than regular folks getting on ChatGPT and having it do work reports
LOL.

EliteZags said:

---

so Palantir is getting loads of 8 and 9 figure contracts left and right for something that may become useful a few years out?

---

I am not saying that's the case necessarily. What they're doing for DoD is likely one of the specific use cases of utility for AI, which I'd presume is much more related to navigation / systems monitoring. However, look no further than the gigantic VA / DoD contract with Cerner (now Oracle) for an example of a technology trainwreck that no one saw coming except those of us intimately familiar with Cerner.

The healthcare sector has been on the doorstep of trying to make use of AI to analyze all the data they collect and still no one has figured that riddle out yet, and I fully expect that to still be the case in 2026 and maybe still in 2028.

Palantir has partnerships with several hospital systems in the US and including the UK govt healthcare system.
In some of the first use cases, they have shown that they can help nurse scheduling, using info to free up beds and streamline the discharge system.

One hospital system developed a real world use case created at one of the bootcamps.
This product was noted by the hospital admin. that in one department, of one hospital where this was deployed would be an annualized savings of over a million dollars.....this was for one deptartment at one hospital,,,,,not across the whole system.

Palantir also just partnered with Oracle- who moved their hq to Nashville to be closer to the center of the hospital system hub. Reading between the lines I feel like behind govt, healthcare it the primary target for some big AI investment and real innovation in the space.

I agree that there is plenty of reason to think AI will improve our healthcare inefficiencies. However, no significant progress has been made in doing so across most of that industry. It has been a lot of over-promising and under-delivering and I say that as an insider from one of the larger systems (Ascension) and as someone who worked extensively with both Athena and Cerner on one of their more "groundbreaking" data integration initiatives that was a complete flop so far.

Untold billions have been spent chasing elusive goals set by overly optimistic hospital administrators who saw a presentation and jumped to conclusions that what they just saw is the next greatest solution to all of their problems

Still way early. The compute power is not there yet.

I think AI mania is propping up the entire market. It's still in a very early phase though. A lot of companies are buying/leasing AI chips without having much in the way of software to support the expenditure. I think the current AI chips and hardware will expire in usefulness before any decent AI software makes it to market. Overall, it's just growing pains...

All can be true…

AI is extremely useful in a lot of situations already

AI is overhyped in many cases

AI is tremendously valuable

Some AI stocks are dramatically overvalued

Some AI stocks are dramatically undervalued

right and basing the trajectory off performance of image generators / coworkers of someone's wife is like thinking you can land a plane cus you watched Top Gun last night

RoyVal said:

---

Four Seasons Landscaping said:

---

OP sounds like a very new headline focused approach to AI.

The people that actually know how to use AI aren't really impacted by the "flubs" he mentions.

---

Agreed. AI is so much more than regular folks getting on ChatGPT and having it do work reports
LOL.

---

Explain what it is then.