Dearest Rachel –
Compared to the last couple of days, there’s not a whole lot going on today. I mean, there’s enough to keep me occupied, even busy, for that matter, but nothing particularly out of the ordinary that would be worth telling you about, and which you couldn’t point out that I’d told you about from a previous day out of the last eighteen hundred or so. Now, it’s true that, in life, there are many days that are much like one another, to the point where it sometimes gets hard to tell one apart from another, but that doesn’t make for a particularly interesting story, so let’s not repeat an ordinary ‘day in the life.’
Instead, assuming you’re willing to bear with me, I’m going to expand on a reference I made in passing just yesterday about my growing conversations with A.I. chatbots. It’s wild to realize that, three years ago, these were so new and unusual that it seemed that the little interactions I was having with them put me on the bleeding edge of technology. Nowadays, it’s as if every customer-facing tech company has their own version, each with its little quirks and such, but all of them trying to replicate a human interaction. Microsoft call theirs CoPilot, which uses the latest ChatGPT architecture (I don’t know if they own OpenAI outright, or if this a partnership); Google has developed Gemini, which has come a long way from when it depicted nothing but minorities in its attempts at artwork; and Twitter (well, they call it “X” now) borrows from Heinlein and calls theirs Grok. There are others out there, such as Claude and Perplexity, but you have to seek those out deliberately, rather than having them packaged with a software you already use. The competition is fierce, with everybody trying to either be the best, or at least to carve its own niche in the ecosystem.
Each of them that I’ve interacted with seems to converse in a fairly human-like way; really, the only way they screw up the Turing test these days is in the fact that their response speed is so fast and so thorough that you know you’re communicating with a machine rather than a human. And with more processing power being devoted to the programs all the time, it will ever be thus going forward. It’s what we want as consumers; results as quick as possible, delivered in a friendly (bordering on obsequious, in some cases) manner.
That friendliness is actually a bit of a drawback; I’ve seen a few instructional videos as to how to get honest feedback from these chatbots, as they tend to agree too much with their users, in order to get them coming back for more conversation. That’s fine if you’re wanting to just chat with them, or request information (which, to be fair, has been most of what I’ve used them for up until now), but if you have a decision to make, sometimes you need to be told your idea is a bad one, and they aren’t good at that. They have to be specifically prompted to be suitably honest, as they are by nature loathe to tell their users that their idea might be unworkable. Brutal honesty isn’t part of their natural programming, as the assumption seems to be that it would turn off users for being unfriendly. But interestingly, there is supposedly a sliding scale among the various architectures as to which is more reliable in terms of feedback, and which is a little more sycophantic. It so happens that ChatGPT lands about in the middle of the pack, while Grok tends to be the most blunt about telling its users whether a decision is worth pursuing or not.
So, with regard to making certain purchasing decisions, I’ve made a point to check in with Grok, often asking it to give me a ‘Consumer Reports’-style analysis of the various options of the purchase I’m considering. In terms of the situation I told you about yesterday, it’s given me a half-dozen options that are comparable (and in most cases, preferable) to either the Explorer or the Edge – the Ford models that Dad recommends I look into to replace the Taurus, assuming it’s the size of vehicle I still want. It is nice for schlepping the gang around, and for bringing everything along on a road trip, but it’s probably more than I need for everyday use (which, ideally, is where the Twike and the tuk-tuk might come in).
But while Grok offers an enthusiastic, if qualified endorsement (due to it still being in the planning as opposed to actually being on the road as of the moment) of the Twike – saying that “for what it’s worth, [it] looks like a solid, eco-friendly choice for urban/suburban zipping once it hits production” – and presented me with a list of a handful of models to consider, it made it clear that tuk-tuks, while technically street legal, aren’t particularly optimal for tooling around in here in the suburbs. Essentially, they’re like exotic-looking golf carts, with a speed to match – which is particularly problematic on the street where we live, since the posted speed limit is 35 mph, right at the top speed of these machines. And I hardly have to point out that not everybody driving down our street is going the posted limit; to hit what we used to refer to as ‘the natural rate of speed’ as we explained to Daniel when teaching him how to drive in traffic, I’d have to be able to start out at a speed that exceeds most of their specifications.
Moreover, the gas-powered ones, while having a little bit more in the way of pep – and thus almost passable in our local traffic – have considerably rougher rides, and would require me to test for a motorcycle license. While I imagine I could handle the process, it’s enough to make me reconsider looking into one of these things. So imagine that; an AI chatbot talked me out of a purchase. Which, you’d think, would be the opposite of what its makers would have it do, as you can bet that these things will have ads in them to monetize them sooner rather than later. But at least for now, I can get a reality check from one of these things, and if need be, be brought back down to earth on one or another of my outlandish plans. I have to say, I rather appreciate that; now, if only I could narrow down my picks for an actual auto (and if I could be allowed to look into a Kia or a Honda as opposed to a Ford – I always knew I was the black sheep of my family).
Anyway, I need to get on with the rest of my (relatively boring) day, honey. Keep an eye on me, and wish me luck. I’m going to need it.
Letters to Rachel 