What I find particularly interesting about this philosophical approach is how much it comes across like nothing so much as just searching for a plausible excuse for intellectual laziness.

Real introspection can be challenging and lead to some uncomfortable realizations about yourself, so it's understandable why someone might want to shy away from it. But rather than just admit that as a shortcoming and an opportunity for personal improvement, because admitting the need for personal improvement is also challenging, it's easy to concoct a reality where lack of introspection isn't just OK but actually the preferable alternative.

The person you're replying to was talking about "virtue", which I'd argue is an entirely different thing than "morality". Virtues are traits that can help people be better humans, while morals are rules separating right and wrong. Things like courage, patience, introspection, kindness, etc, are all virtues, while morals tend to be much more of the Ten Commandments variety.

I don't think anyone has ever killed people or gone to war for a virtue, but they certainly have to enforce their moral code on others. Probably the worst combination is people with strong moral beliefs but few virtues, since their lack of virtue both fails to temper their moral fury and poorly guides the moral determinations they get so fired up about.

> Virtues are traits that can help people be better humans, while morals are rules separating right and wrong.

Insofar as there is a difference at all, (what people see as) virtues are generalizations of (what they hold to be) moral rules or, viewed another way, moral rules are the operationalizations that give concrete meaning to the platitudes of virtues. They are inseparable, even if notionally distinct categories.

We moved from the mainframe era to desktops and smaller servers because computers got fast enough to do what we needed them to do locally. Centralized computing resources are still vastly more powerful than what's under your desk or in a laptop, but it doesn't matter because people generally don't need that much power for their daily tasks.

The problem with AI is that it's not obvious what the upper limit of capability demand might be. And until or if we get there, there will always be demand for the more capable models that run on centralized computing resources. Even if at some point I'm able to run a model on my local desktop that's equivalent to current Claude Opus, if what Anthropic is offering as a service is significantly better in a way that matters to my use case, I will still want to use the SaaS one.

> Even if at some point I'm able to run a model on my local desktop that's equivalent to current Claude Opus, if what Anthropic is offering as a service is significantly better in a way that matters to my use case, I will still want to use the SaaS one.

Only if it's competitively priced. You wouldn't want to use the SaaS if the breakeven in investment on local instances is a matter of months.

Right now people are shelling out for Claude Code and similar because for $200/m they can consume $10k/m of tokens. If you were actually paying $10k/m, than it makes sense to splurge $20k-$30k for a local instance.

The underlying advantage of local inference is that you're repurposing your existing hardware for free. You don't need your token spend to pay a share of the capex cost for datacenters that are large enough to draw gigawatts in power, you can just pay for your own energy use. Even though the raw energy cost per operation will probably be higher for local inference, the overall savings in hardware costs can still be quite real.

I'm not sure that's infinitely true as long as AI costs to the user are proportional to the cost it takes to run the model. Even if user costs are heavily subsidized by investment, as long as they are non-zero and go up when models cost more, there will be at least some pressure for cheaper models and not just more capable ones and that pressure will go up with costs. AI is a crazy industry, but it's not totally immune to the law of supply and demand.

The real question though is how close are we to the point where the pressure is more for efficiency rather than capability. Anecdotally I think it's a ways off. Right now the general vibe I get is that people feel AI is very impressive for how cheap it is to use, which suggests to me that a lot of users would be very willing to pay more for more capable models. So the tipping point where AI hardware demand might slow down seems a ways off.

It has yet to be seen when hyperscalers will change their tune.

I wonder how much is truly preference at this point rather than societal inertia. Actual walkable cities with good public transit are incredibly rare in the US and as a result tend to be very expensive (which itself should tell you something about demand). Most Americans have no choice but to live in an area that requires a car for daily life. I'm sure there are plenty of people who would choose the car dependent lifestyle even if given the choice not to, but the demand for alternatives is probably higher than you think.

I also think there's a very reasonable middle ground where people can still practically have and use a car but it's not required literally every time you leave your house. Personally I think giving up my car would be a bridge too far since I like road trips and drives out to hiking areas and things like that, but I also find it unfortunate that there are limited options of affordable places to live where I don't need a car to do everything.

I see you here - you make very good points. It just seems like economically it’s too hard to manifest more of those pleasant, walkable cities and neighborhoods into existence in a way that they aren’t just as costly to live in as the existing ones. You can’t just build Williamsburg on the next available spot of land next to the last suburb, because no one will want to live there without a car when there isn’t a subway right in the neighborhood to take them to Manhattan in 25 minutes.

So, they just aren’t making many net new walkable cities or converting previously car-dependent ones into walkable paradises. The only newly-built ones are built on insanely expensive land (because of the proximity to great transit and/or very high paying jobs), so they’re really only feasible for people with at least $250k annual incomes, which isn’t most people. I think a lot of people know all this instinctively and therefore are against even talking about it because they see it as an unsolvable problem.

Yeah I see your point. I think only NYC and Chicago are truly walkable cities that we have in the US and yes different people have different preferences. I find that a lot of people in the US do not want to live the walkable city lifestyle once they hit their mid late 30s but I’m sure they are plenty of people to do and it would be nice for them to have more options to move to in that case. I loved living in Manhattan for 10 years when I was in my 20s but now in my 30s, I would not want to go back and do that again. I really appreciate having a lot of property and the freedom that comes with having a car and being able to just go anywhere you want and not having to rely on fixed public transit routes. Usually when you say something like this people come back with you can just rent a car when you need one, but that added inconvenience of having to rent a car means you’re just not going to do it the vast majority of the time.

Why do you think anybody was operating under the assumption that this was free? But keeping your car topped up now is hardly free either, especially lately, so the question is really about cost comparison. And that's before you get into any externality costs.

> so the question is really about cost comparison

Yes, and I was talking specifically about the cost of this difference.

Why? The vast majority of cars spend most of the day stationary. I'd even venture to say most cars spend most of the day stationary in the same spot. If that spot has charging, slow or not, it would likely cover the daily energy used by that vehicle. Aside from road trips, that literally sounds like the perfect charging setup to cover most vehicle use-cases.

I'll take the possibly controversial position that WireGuard's opinionated approach to cryptographic choices without the option for negotiation was indeed the right call, but it would have been a better and even more successful protocol if it used FIPS compliant cryptography.

Taking the DJB crypto path gave Wireguard some subtle advantages to implementation ease-of-use that are almost entirely overshadowed by the difficultly in building a new, secure cryptographic protocol from scratch regardless of what algorithms you're using. The tradeoff was that there are plenty of places it will never be used due to standards compliance requirements which as you point out also has significant implications for efficiency in hardware.

Wireguard is cool. I think very little of that coolness has to do with the DJB vs NIST cryptographic choices. And taking the DJB path unnecessarily limited the impact of its coolness at least for now.

I think cryptography engineers increasingly agree with this take, but it's also a different world: it would be straightforward to do XAES and modern P-curve implementations (now that they've been worked out with complete addition and stuff like that) now, but that was less the case when WireGuard was first published.

You're right that it seems close enough, but only as long as we're talking about the time it takes a single vehicle to "fill up". Taking 2 minutes vs 8 minutes to fill up your car doesn't matter to you personally, but it is a significant difference to those installing and operating the fill up infrastructure since it takes 4 times as many charging points as it does gas pumps to serve the same volume of customers in a given period of time.

In a lot of cases that probably won't matter since chargers can be installed in more places than gas pumps and for gas stations that serve mainly local customers I'd expect demand for an equivalent charging station to be lower since some people will charge at home at least some of the time. But things like highway rest stops could be more of a challenge since you'd expect customer volume for EV charging to be similar to the demand for fuel so you'll need more charging stations at each stop to handle the increased time it takes each customer.

Unless you're independently wealthy, even the most traditional two parent nuclear family with a married mom and dad requires at least one parent to spend a significant portion of their day away from their children earning a living. I'm not sure why you would so strongly object to both parents doing that when you would presumably not object to just one doing the same.

WFH kind of solves that. Instead of taking coffee breaks or going to lunch with the colleagues presumably one of the working spouses can spend those breaks with the family. Now, I am looking at this problem from the European perspective where one of the spouses receive childcare payments and is allowed not to go to work for 1-2 years. I know about American system (12 weeks off) and it's beyond insane to me.