World Finally Realizing AI Can't Fix Everything

Saturday, September 27, 2025

Graham Templeton: It's become trendy to compare the current rush of investment in AI to the "dot com" bubble of the past, but the comparison just keeps getting more apt. There's obviously the almost absurd level of investment that's hitting tech, with the broad tech sector now accounting for 34% of the S&P 500's market cap, but there's more as well.

Posted by lamplighter at 08:32 PM | 2 COMMENTS | permalink | Comment on This Entry |

Alternate links: Google News | Twitter

MIT recently published a study showing that companies are experiencing reliable operational slowdowns due to the integration of AI. The leading cause of this decrease in efficiency is that many employees who use AI to complete most or all of a job-related task actually pass the work on to someone else (including, potentially, their future selves), as someone has to correct the AI's shoddy output. This forces the next person in the chain to deal with the inherently low quality of the work.

Comments

Admin's note: Participants in this discussion must follow the site's moderation policy. Profanity will be filtered. Abusive conduct is not allowed.

More from the OpEd ...

... For one, there are the ridiculous over-promises and the credulous investment managers who gobble them up, leading to inevitable backlash. It is no longer controversial for even respectable legacy financial publications to refer to the current situation as a bubble or to sound the alarm on the economic dangers that a bubble could pose.

We can't know yet whether this particular investment bubble will burst or merely deflate, but one thing seems clear: We don't have very long to wait before we find out. The past few weeks have seen a litany of negative headlines about AI, and for once, it's not just alarmist worries about unemployment or Skynet super-intelligence.

This time, people are questioning the specific claims that have historically helped certain people overlook AI's potential dangers. The two primary such claims to value are, in order of importance: It will make money, and it will help people. ...

#1 | Posted by LampLighter at 2025-09-25 12:24 AM | Reply

If you want to have a laugh, try this out on several different AI chatbots:

List cities ending in the letter n

Then see what the AI spits out. Some are better than others, but every one I've tried so far will include cities that definitely are incorrect.

THEN carry on a conversation about why it included the incorrect cities and press hard about it. You can even get them to say a correct city is incorrect.

Ultimately I got a couple of them to admit they included incorrect cities because it was more important to return results that would appease me rather than being correct.

And the thing is, it's such an easy thing to verify that a city ends in a particular letter. And yeah, I've pointed out to the chatbots how easy it is to verify and have them try again only to have most of them still return erroneous answers.

#2 | Posted by Idependant97 at 2025-09-25 01:53 PM | Reply