More generically, our species' Achilles heel is our inability to factor in the long-term cost of negative externalities when evaluating processes that yield short-term positive results.
This. From simple personal choices to the marker economy and politics. With games we're introduced to cheat codes pretty early in our lives. Some people outgrow them, some don't. Too bad our systems encourage their use, whether it's a time-to-market thing, cutting costs, or the next election.
Counterargument. The author is primarily looking at AI trend lines. Let's say our industry continues moving along alternate, equally compelling, trend lines: increasing global volatility, chaos in the energy markets, growing likelihood of great power conflict this century, climate collapse, mass migration, societal unrest, yada yada.
What happens to all of these AI-native companies if the AI bubble is not able to survive in these conditions? If your current development process is built on the metabolic equivalent of 400kg of leaves per day[0], then when the allegorical asteroid hits, you're going to be outperformed by smaller, nimbler companies with much lower resource requirements. Those companies may be better suited for survival in hostile macro conditions.
In other words, I think a lot of companies believe that they're trimming their metabolic fat by replacing engineers with AI. Lower salary costs! But at the same time, they're also increasing their reliance on brittle energy infrastructure that may not survive this century. (Not to mention the brittleness of the semiconductor fabrication pipeline, RAM availability, etc)
Predicting the future isn't about being right tomorrow, it's about selling you something today. - read that somewhere
Folks using AI aren't interested in the future, they are interested in buying today and maximizing profits today. If something goes wrong tomorrow, then that's when the problems are dealt with: tomorrow.
AI is an incredibly fragile technology, as you say, it's depended on so many things going right, amazing stuff that it works at all. That fragility includes price, once that goes up and developer price goes down, the winds of change might blow again.
AI also forces folks to be online to code, without being online, companies cannot extend their products. Git was the first version (open source) control system that worked offline. We're literally turning back the hands of time with AI.
AI is another vendor lock-in with the big providers being the sole key-holders to the gates of coding heaven. Folks are blindly running into the hands of vendors who will raise prices as soon as their investors demand their money back.
AI is "improving" code bases to make subtle errors and edge cases harder to detect debugging without using AI will be impossible. Will a human developer actually be able to understand a code base that has been coded up by an AI? That's a problem for tomorrow, today we're making the profits and pumping up the shareholder value.
AI prompts are depended on versions of LLMs - change the LLM and the prompt might will generate different code. Upgrade LLMs or change prompts and suddenly generated code degrades without warning. But prompts are single-use one-way technology: once the generated code is in the code base, there is no need for the prompt - so that's non-issue, only for auditors.
Having come from levers, to punch cards, to transistors, to keyboards, to mice and finally AI, programming has fundamentally forgotten there is a second dimension. Most fields have moved to visual representation of data - graphs, photos, images, plans etc. Programming is fundamentally a single dimension activity with lines and lines of algorithmic code. Hard to understand and harder to visualize (see UML). Now AI comes along and entrenches this dependency on text-based programming, as if the keyboard is the single most (and only) important tool for programming.
It's a lack of imagine of exploring alternatives for programming that has lead us here. Having non-understandable AI tools generating subtly failing code that we blindly deploy to our servers is not an approach that promises look term stability.
This isn't true in the broad sense you've used. It's true that most people don't have the hardware to run the bleeding-edge foundation models, but with a modest Macbook you can run very capable local models now (at least capable for coding, where my experience is).
Here I was talking of the AI vendors - they specifically provide inferior models for local usage while offering the "insanely" good models only online.
AI can be run locally but with the growth of agent factories, this is going to be less and less possible if you want to keep up with the Jones.
> AI is "improving" code bases to make subtle errors and edge cases harder to detect debugging without using AI will be impossible. Will a human developer actually be able to understand a code base that has been coded up by an AI?
Huh? It’s just code that you can read. Why do you think the code will be impregnable by a team of human minds?
Because code does not include the thought processes that went into creating the final code. Take a second and have a look at the Linux kernel code base and get into that. It's surprising how some code only make sense if you understand the bigger picture.
So it will be with AI code that has just been generated and blindly added to the code base. It makes everything work but sometimes, perhaps not always, the devil lies in the details.
Take any book, open it up to a random middle section, read it. I can read the words but I don't understand the story. And so it is with code.
So, how does aj engineer new to a code base add new features? They read the code base, understand the architecture and structure and make changes. An agent can do the same.
> Right now there are companies which hire software devs or data scientists to just solve a bunch of random problems so that they can generate training data for an LLM model.
> Which begs questions about whether closed source will provide any protection (it doesn't appear so, given how able AI tools already are at disassembly?)
Disassembly implies that you're still distributing binaries, which isn't the case for web-based services. Of course, these models can still likely find vulnerabilities in closed-source websites, but probably not to the same degree, especially if you're trying to minimize your dependency footprint.
If you haven't tortured yourself on the Devil's Corkscrew switchbacks on the Bright Angel Trail at the Grand Canyon, on the hottest day of summer, have you really National Parked??
All joking aside, I disagree with the author regarding the Grand Canyon. Havasupai Gardens -- the verdant oasis at the bottom of the canyon, where you can camp and recharge -- is one of my favorite places I've camped. There are areas for wading and swimming, and the sounds of the night creatures is eerily beautiful.
We've only hiked a little bit down the canyon when we were last there (which was back in 2009, sadly), but I also strongly disagree with the author. It may not be the longest, deepest or widest canyon in the world, but if you actually go there and take a look at it, you'll see for yourself why it's called the Grand Canyon. Then again, if this article keeps some easily bored/jaded people away from the national parks, all the better for the rest of us!
I hiked and camped for three days in the Grand Canyon about 10 years ago and it was one of the high points of my life. The sheer beauty of the massive space was transformative in the way I see the world. How small we are, how big it was. I know statements like this seem overly dramatic but my time there is still something I replay in my head.
So when people dunk on the GC, I always just tell myself how lucky I was that my experience was so wonderful.
The problem with the Canyon is that (1) what's really interesting is seeing the whole thing top to bottom and (2) the really spectacular views are when you're partway down, with the best locations depending on the time of day. However, because it's such a challenging hike, most people don't get those experiences. FWIW, Havasupai Gardens is only about halfway down, but because it's the last water tap before the river, it's kind of the last somewhat readily accessible point.
I half agree with the author's take on Yosemite. The valley really is ridiculously overcrowded but the view also is amazing. The situation here is a little bit like the Grand Canyon in that there's lots of amazing stuff but it's more work to get there, mostly in terms of a lot of time on foot. The good news about the Sierras is that except for the amazing big wall stuff in the Yosemite Valley, Sequioa Kings Canyon is basically just as nice and the traffic situation is much better.
We drove to the Grand Canyon in a rental car from Vegas. "We" that is my old parents and I. We don't gamble, drink, or do hookers, so we didn't really like Vegas very much. (Stop! Hold your horses. It's a joke. I know there's other things you can do in Las Vegas. It's a joke.)
We arrived at the canyon before sunset and the setting sun hit it so beautifully from above. The colors were incredible.
We were gonna just spend the night and drive back first thing in the morning. But then we said to ourselves, we don't like Vegas, the Grand Canyon looked awesome. So we went back the next morning and spend more time there, just driving along the rim and stopping every now and then to let the impressions sink in.
With the sun on the opposite side now than the night before, it looked totally different.
This was all part of a longer road trip, and the Grand Canyon was definitely one of the highlights.
That's similar to my experience. I used to go to LV for the DefCon conference, and a classic gaming expo one year. But I'd go for a week, and also drive around the area, including over to Death Valley. I eventually started spending most of my time on Death Valley and only visiting LV briefly.
As someone who doesn’t hike or camp, I didn’t really enjoy the Grand Canyon.
When I was 18 my dad thought it would be cool to take me there for a week. I suggested we just stop there briefly, then go skiing in Colorado, which he accepted.
We got to the Grand Canyon and it was so big that it didn’t even look real. I was pretty much done after 30 minutes, so was my dad. He was glad I suggested the Colorado option, as he wasn’t sure what we’d do there all week.
I’m glad I saw it, so I’ll never wonder, but I’m also glad I didn’t try and make it into more than it was, as a non-camper.
I like hiking but absolutely not camping. I had a similar experience. IMO the canyon is too wide for its width which creates a similar effect to it just being a valley. I found it much more spectacular to stand on pretty much any mountain in the Alps.
Funny enough driving up to the grand canyon there are some arms of it that in essence narrow, deep cracks in the ground. I absolutely loved those!
I-15 out of Vegas hits a small part of Arizona on the way to Utah, and the freeway runs up a river gorge - it's an amazing drive and in no way comparable to the Grand Canyon - but it has its own charm.
They could just be writing for themselves, or their friends, or for people with the patience to read. You are making assumptions about how badly they want to reach your particular eyeballs. They might not care about trying to win over people with a minimal attention span as much as you think they do.
What makes you think your comment was worth reading?
More generically, our species' Achilles heel is our inability to factor in the long-term cost of negative externalities when evaluating processes that yield short-term positive results.
reply