the_snooze

Proper engineering requires observing and working with real-world constraints. Ignoring constraints is a form of magical thinking.

One real-world constraint is that end-users won't always do what you intend for them to with the thing you engineered.

>Decapitating the regime and degrading their military capabilities was the primary objective.

Those are tactical objectives, not strategic aims. The US is very good at winning tactically, but losing strategically. This is yet another example.

Bringing on new fabs takes many years and billions of dollars. You're exposing yourself to a lot of risk if you build now and find that the gold rush is over by the time your new capacity is online.

That kind of inevitability rhetoric is a big reason why people dislike AI. It's an impressive technology sure, but impressive doesn't automatically mean operational. It's got serious issues with reliability today, and appealing to some possible future state is less rigorious engineering and more unfalsifiable magical thinking.

We're going back to the mainframe model. Client-side general-purpose computing is an impediment to recurring subscription revenue and vendor lock-in.

>Absolutely agreed. It feels like tech companies forgot that they are supposed to add value to users.

They lost the plot long ago. They're firmly in extraction mode now: how much value can they get from end-users?

I'm of the same mind. Money is simply a means to an end. The workplace is just somewhere to exchange your time and skills for money, which you use to pursue priorities outside of that.

No job will ever tell you you've done enough and so you should do something else. If you don't make that call yourself, whose life are you actually living?

You don't. You accept that social bandwidth is a real-world constraint that you work with, not magic away. That's real engineering.

This is why I asked:

>And why not take the alternative approach of identifying the subset of people who have indeed found solid uses and spread their best practices around?

A bottom-up approach has a far better chance of finding those particularly good use cases, and if you lean on the people how found those fits, they're more persuasive than top-down edicts. They actually know what they're talking about. If the point is to leverage AI for better work outcomes, someone with your experience is far more valuable than "here's a dashboard, make the number go up," which seems to be what's going on at Amazon.

My questions for that approach are: Why treat AI as a special technology that needs enterprise-scale exploration to come up with a useful application? And why not take the alternative approach of identifying the subset of people who have indeed found solid uses and spread their best practices around?

The top-down approach to encouraging (mandating?) AI usage strikes me as infantilizing to the workers, who are perfectly capable of choosing which tools they use and when.

>I do agree that it is still a tool to keep in touch with some people, yes, but… Are there no alternatives that require a lower price to pay?

The sustainable alternative I've arrived at is private group chats with notifications disabled. Engagement is entirely on my own terms (i.e., I have to manually open the thing to see any updates), and there's no formalized moderation or amplification because it's human-scale enough to let shared social norms govern it.

These benefits make sense up-front, but we have 15+ years of direct experience with "democratizing" technology that ultimately end up ensnaring people. Unless you get these benefits from a local model, you're establishing a deep dependency on an AI service provider whose interests may not necessarily align with yours.

Too much of modern consumer-facing software think they're the ends, not the means.

altek has been given a number of off-ramps and alternatives to proceed. His continued resistance to take those isn't a sign of naivete, it's a sign of bad faith.

I gave up on Spotify when they did their push into podcasts and audiobooks. It became clear that they weren't really interested in serving their core customer base of people who just want to listen to music.

This is really the heart of it. Understand what matters to you, and align your finances (and decisions as a whole) with that. If you optimize for some external metric without examining your own values, you're basically adrift.

I wouldn't necessarily say "idealistic," but certainly constrained. Microsoft has always been scummy in one form or another, but always-on internet connectivity has allowed them to be scummy in persistent ways long after your purchase of their product. It's a serious money-maker, but I think that explosive growth has bred a whole generation of tech "professionals" these days that think more like Wall Street bros than sober engineers: make line go up, damn the consequences.

They got way over their skis on this one. There's a difference between "impressive" tech vs. "operational" tech. That difference usually boils down to prioritizing engineering rigor over marketing.

Exactly, these are highly-paid professionals with very broadly applicable skills. They have the means to uphold professional values.

>It's restricted because it's genuinely good at finding vulnerabilities, and employees felt that it's not a good idea to give this capability to everyone without letting defenders front-run.

It's a possibility, but it doesn't eliminate the possibility that it's hype. If these claims were indeed serious, they would submit it for independent analysis somewhere.

This isn't some crazy process. Defense contractors are required to submit their systems (secret sauce and all) for operational test and evaluation before they're fielded.

>Even searching for specific topics is hard.

It's the tyranny of the marginal user. How I wish YouTube (and generally other platforms for user-generated content) would have fine-grained search and filtering controls that let me specify exactly what I want, no recommendation algorithm trying to guess what I actually meant. But such a feature won't attract and retain the least interested people, so we'll never get it.

Yeah I'm of the same view. If you're driving, the least you can do is take the responsibility seriously and give your surroundings proper attention.

An automated fine is the least painful way to enforce that.

> being able to be "present" without the time cost can mean (possibly significantly more) presence at the same cost.

This is magical thinking. "Presence" and "time cost" are inextricably linked. You can't have one without the other.

When you use AI to decouple them, you're telling your audience/colleagues/attend that you want them to listen to you but not the other way around.

That sounds like straight up scammer behavior. "Yes, this is Microsoft calling. We need to confirm your info with the local authorities."

>I stress test commercially deployed LLMs like Gemini and Claude with trivial tasks

I did exactly what I said I did. I'm using these systems the way they're designed and advertised. I'm following the happy path with tasks that are small, trivial, and easy to check. This is the charitable approach. Yet the system creaks under the lightest load. If Google wants to put on a better show with stronger models, then they should make those the default.

You don't need to make excuses for shoddy engineering from multi-billion dollar corporations. And you're quite welcome to run the same prompt on ChatGPT and evaluate it on your own time.

I did this in one attempt just now: https://gemini.google.com/share/b4e016be1f69

#8 has an incorrect answer (3 appearances according to Gemini, 2 according to reality https://en.wikipedia.org/wiki/Bowl_championship_series#BCS_a...)

So it works well 95% of the time for literally a trivial use case. Imagine if any other tech tool had that kind of reliability: `ls` displays 95% of your files, your phone successfully sends and receives 95% of text messages, or Microsoft Word saving 95% of the characters you typed in. That's just not acceptable.

There's nothing in these models that say its purpose is software development. Their design and affordances scream out "use me for anything." The marketing certainly matches that, so do the UIs, so do the behaviors. So I take them at their word, and I see that failure modes are shockingly common even under regular use. I'm not out to break these things at all. I'm being as charitable and empirical as I can reasonably be.

If the purpose is indeed software development with review, then there's nothing stopping multi-billion dollar companies from putting friction into these sytems to direct users towards where the system is at its strongest.

Two things can be true at the same time: The technology has improved, and the technology in its current state still isn't fit for purpose.

I stress test commercially deployed LLMs like Gemini and Claude with trivial tasks: sports trivia, fixing recipes, explaining board game rules, etc. It works well like 95% of the time. That's fine for inconsequential things. But you'd have to be deeply irresponsible to accept that kind of error rate on things that actually matter.

The most intellectually honest way to evaluate these things is how they behave now on real tasks. Not with some unfalsifiable appeal to the future of "oh, they'll fix it."

That's how I do it too. I'll tap bell once (and let the ring sustain) when I'm about ~5 seconds from overtaking them so people know there's something coming up behind them, and the sustained sound tells them how fast it's coming. This is especially important with runners, who are prone to suddenly take a U-turn if they're at the end of their route.

Pedestrians regularly wave acknowledgement or even say "thank you." Some other cyclists (especially on e-bikes) just blast by with no warning.

I'm not a fan of the "something better" phrasing myself. It's very much anti-systems-thinking.

Engineers should be honest that everything is a tradeoff. For the up-front convenience you get with phone tickets, you impose additional failure modes, dependency chains, and accessibility issues that simply weren't a problem with paper ticketing.

The "phone-ification" of everything will probably bite us in the behind in the future, just like the buildout of out car-centric environments does now.