3 mostly annoying AI trends we see in companies right now

00:00:00: When it comes to companies using AI, we are probably here when it comes the current hype about AI.

00:00:11: I'm not sure if reached maximum or top yet but definitely at a very high point now.

00:00:19: obviously I'm kind of no exception.

00:00:45: that I enjoyed writing code before we had AI, but it is what it is and of course nothing stopping me from still writing code by hand.

00:00:56: But if you are doing serious development work... If your building software... Yeah!

00:01:03: AI can make you more productive and not using it as a valid choice.. ...but in most cases probably not the right one if you're doing it professionally.

00:01:14: That is my opinion at least!

00:01:16: Now that does NOT change though, that we are really a top point regarding hype in companies and maybe will go to even higher heights now.

00:01:31: of course when I say AI IN COMPANIES i'm fully aware there's technically not an endless amount of companies, but there are many many companies all over the world and The use of AI is not evenly distributed.

00:01:46: And I'm also mainly talking about tech-based software oriented companies here.

00:01:52: But when we talk or look at the uses in companies There's a couple Interesting and kind of worrying trends we have to talk about.

00:02:06: We have to look at one big trend, we saw over the last couple of weeks or we heard About The Last Couple Of Weeks is for example token maxing.

00:02:19: Token Maxing of course means that For some companies more AI use is better.

00:02:26: That's it.

00:02:28: There are Reportedly, companies that had or have internal leaderboards track how many tokens their software developers and employees in general spend on AI.

00:02:46: And well, that is a bit like tracking just the lines of code you generate as a measure off how good your doing.

00:02:55: It's a bad measure and I probably don't need to tell you it should be pretty obvious Measure to measure how productive you are.

00:03:08: I mean for one It's easy-to-game.

00:03:10: You can just send pointless prompts to the AI or have it do pointless meaningless work And that allows you to max out your tokens, but even if you're not doing that Which developer is probably better?

00:03:26: The one that puts some thought into a problem that analyzes the result DAI gives them or suggests and then analyzes to code, OR the vibe coder just prompts away and produces tons of tokens for some companies.

00:03:47: It seems number two!

00:03:49: For me it's definitely number one.

00:03:52: I think the magic thing about DAI and where you really can get a lot of use out it right now is if you do combine your expertise with The advantages of AI which, of course are that.

00:04:06: You can move fast That you have an infinitely patient Mentor Where?

00:04:12: Can ask questions that you kind of combined two brains though you should really trust your brain A bit more in most situations.

00:04:21: Yeah, you can get advantages out of AI but token maxing is probably not it.

00:04:27: But there's a trend we see or heard about in some companies though now that they've got public, some of these companies like Meta for example already announced were getting rid off the leaderboard and strong incentive to max out tokens….

00:04:47: another trend?

00:04:50: I saw more and more over the last week is that there are companies like McKinsey, who really want to push the narrative of AI agents as employees.

00:05:04: The McKinsee CEO here said they would have about sixty thousand employees where twenty five thousands of them are AI agents.

00:05:16: So AI Agents as workforce is something I've also seen here and there.

00:05:22: And yeah, of course that's one way to look at it… but then again….

00:05:35: Most all AI agents are pretty specialized regarding what they can do, whereas humans tend to be more versatile in what they Can Do and What They Can Be Taught.

00:05:49: So I'm not sure if that comparison really makes a lot of sense... Who would have thought of calling your backup script?

00:05:58: An employee that you might have had in two thousand eighteen already, right?

00:06:02: So we had automations.

00:06:03: We had workflows before AI and we did.

00:06:07: That makes a lot of sense.

00:06:09: obviously we have automated deploy processes backup processors web scraper's data analysis Processes.

00:06:15: we have all kinds of automations And we have them for ten twenty years or longer.

00:06:21: But now it's the AI agents that we call parts off The workforce.

00:06:26: and don't get me wrong You can do useful stuff with AI agents.

00:06:32: I'm personally not fully on board with the Open Claw hype because i still am NOT able to get a lot of useful usage out it, but i am using Cloudcode and Codex for coding... Also, for more than that I am using the Pi Agent which is an open source independent AI agent you can use with your chat GPT subscription.

00:07:13: Is that an employee?

00:07:15: I'm not so sure though.

00:07:16: It's a useful tool for sure and i wouldn't even know how to count that.

00:07:22: if i have two employees, each of them is using this agent.

00:07:26: are That then too additional Employees or the agent still just one employee used by multiple other employees?

00:07:43: be the company that knows how to do AI and is on trend.

00:08:08: eyesaw is kind of related to the token maxing, I guess it's mandated use now.

00:08:14: with that however i don't mean That you are required To Use AI as much As possible but You Are Forced to use It Often, or in some cases also specific models.

00:08:28: Or agents.

00:08:29: so that is something That goes hand-in-hand with that mandated use In in some companies out there at least and this Is a point I do get to some degree because i Do understand?

00:08:43: But as A company you want To make sure that your employees Experiment With This new technology And try to find Out where it's useful.

00:08:52: Because I think there is a decent amount of people that used ChatGPT, the free version.

00:09:02: A couple months back or anything like that where they use it here and their end.

00:09:06: It's okay but doesn't seem super impressive.

00:09:10: And especially amongst normal people who are not in this tech bubble The share that are in the bucket, don't regularly use cutting edge models is likely very high.

00:09:25: So I get why companies want to incentivize or push you And that works best of course if you then also get access to the more capable models.

00:09:43: I mean, If i force you to work with an AI model thats two years old That doesn't have any extra capabilities because it's not running in any harness that would give those capabilities... ...that'd be pretty meaningless.

00:09:56: But they do understand why companies want employees To play around with AI.

00:10:01: Now needless to say some companies are definitely pushing too far and there is no sense in forcing employees to do everything with AI.

00:10:11: And I think you should also try to trust your employees when they tell you that a certain task can't be done with AI or at least, it can't get better.

00:10:22: but of course i understand companies can also doubt whether employees really engaged with AI.

00:10:31: And it's easy to push back against the new technology, obviously because many people are afraid.

00:10:36: and how wouldn't you?

00:10:37: With Darryl Amode basically going on a talk show or interview every week telling that vast majority of white collar workers will lose their jobs... I get it!

00:10:49: Many people ARE afraid.

00:10:51: so its easier to pushback against AI.

00:10:53: but i've said this before in other episodes and I strongly believe in it, the only way to deal with this new technology just as with all New Technologies In The Past is To really Embrace It And Try To Get The Most Out Of It.

00:11:09: That Does Not Mean that You Want to Blindly Trusted?

00:11:13: That does also not mean that you want to use it for everything But it means that you should seriously try to use it, see where can help and push the limits.

00:11:24: And especially with AI when everything is changing so rapidly all time... You really also want to re-evaluate regularly every few months or so because things change.

00:11:37: The models changed but more importantly, and I've said that in other episodes too... the harnesses, tools around these models also evolve!

00:11:46: The agents, the agentic harnesses those all evolved and therefore it's very likely, or it's possible that you can do something today with AI.

00:11:58: That You couldn't Do a couple of months back.

00:12:00: so I get this.

00:12:01: but It's Of course clear that some companies are taking This too far and forcing employees to do Something With AI.

00:12:09: that just doesn't work with AI Or where your more efficient without AI?

00:12:14: Of Course is kind of overshooting that target off of incentivizing the workforce To use AI.

00:12:21: But yeah This is where we are.

00:12:24: That's the weird state in which you're right now and I have not really been part of earlier technological revolutions that went through, but i was a child back then when the internet became a thing so... it's very likely pretty normal for things to be rough and weird whenever such a transition or evolution happens, of course with AI.

00:12:59: It is probably especially rough because everything is going so fast!

00:13:03: And due to IT GOING SO FAST AND THE WAY IT WORKS... ...and what people like Dario tell you all the time, its also super frightening.

00:13:13: So yeah All these things combined makes sense that Things are very, very weird.

00:13:20: And again as I mentioned before for me definitely too i had more joy in the coding part.

00:13:28: Before AI was a thing but here we are and yeah Definitely.

00:13:34: right now many companies Are Here?

00:13:37: We'll see if you continue like this or If You find A More Meaningful Reasonable Path Forward Where We Can Really Try To Get The Most Out Of AI Instead of Just Using the most AI without the use part.