Your product has to be a 10x improvement over the incumbent to be competitive.

In AI speak it would be the “extra-bitter” lesson I guess?

You need to add 10x resources to beat a product that’s already solved with mature tech.

Choices are: A. Already happened B. Sometime in 2026 C. 2027 D. 2028 or later E. Never (or practically never)

Comments URL: https://news.ycombinator.com/item?id=47353079

Points: 1

# Comments: 1

If I remember correctly, his talk was about how the world of science-the pure pursuit of truth-and the world of engineering-the practical application of solutions under constraints-had to learn from each other.

In a purely meritocratic sense, all other beings equal a university that provides a diverse faculty and student body will better educate its students than a university that doesn't, all other things remaining equal.

No, the problem is the bad guy the Eastern European countries all want protection from.

I just choose to believe them when they've tried to warn the west about how bad Russia really is.

That's fine and they're lucky to live in a country where they can express any new they wish. Maybe they're right or maybe they're wrong, who knows?!

> From Iraq in 2003 to especially the NATO air campaign in Libya in 2011...

The west is still the preferable choice to support despite these mistakes.

It's a simple question. Do we want to live in a world where Russia achieves their strategic goals or do we prefer to live in a world where NATO achieves their strategic goals?

NATO expansion doesn't happen illegally. It's completely voluntary. It's a defensive alliance meant as a deterrence. And countries in NATO all enjoy much higher standards of living than non-NATO countries. NATO countries all have laws to protect their citizens and they enjoy peace from invasion.

I get that Russia doesn't want that. But my point was so what? I never really denied that issue. Everybody is acting in their best interest. It's just that NATOs interests and values are also the same as my own.

There's no hypocrisy here. There's just a good and bad guy in this case. I don't see the problem here.

Furthermore, if having an interest in something gives the right to use military power to achieve that interest then the argument applies to everyone.

The point about foreign bases in Canada or Mexico gets repeated a lot online, but what is the ultimate point? The USA would not like it, but it's also not a political reality. On the other hand a NATO build out IS a political reality.

So I think rather than focusing purely on what one country wishes it's better to analyze things in terms of what the political realities are and which is better.

In that sense NATO is meant to be a deterrence. Russia doesn't like that. If you ask yourself whose vision of the future is better then the answer is clear. A world of where rule of law is the norm and invasions are deterred is preferable. There has been tremendous peace and prosperity in the EU because of NATO and people have just gotten used to it. They have taken for granted the cost and sacrifice that this peace came from.

However, simply saying that Russia has an interest in not having NATO on their border is almost tautological. Of course they don't want that, but so what. Peace only works if it's enforceable.

If it’s not transformational then this is a bubble and the market will right itself soon after, e.g buying data centers for cheap. LLMs will then exist as a useful but limited tool that becomes profitable with the lower capex.

If it is transformational then we don’t have the societal structure to responsibly incorporate such a shift.

The conservative guess is it won’t be transformational, that the current applications of the tech are useful but not in a way that justifies the capex, and that some version of agents and chat bots will continue to be built out in the future but with a focus on efficiency. Smaller models that require less power to train and run inference that are ubiquitous. Eventually many will run on device.

I guess there’s also another version of the future that’s quasi-transformational. Instead of any massive breakthrough there’s a successful govt coup or regulatory capture. Perfectly functioning normal stuff is then replaced with LLM assisted or augmented versions everywhere. This version is like the emergence of the automobile in the sense that the car fundamentally altered city planning, where and how people live, but often at the expense of public transportation that in hindsight may have sorely been missed.

Basically, safety is "bad things won't happen" and liveness is "good things eventually happen".

This is almost always the way safety is used in a CS context.

Capabilities based programming could come along way though to help with closing that gap.

If we ask who was on the moon before them then the answer is nobody.

I think that’s agreeable. So then what am I talking about? It’s just counting.

I’m going to explain this to whomever is interested, and anyone is free to tell me where I made a mistake, in which case I will thank them for the correction.

When we talk about counting we say we are talking about things like numbers. We also talk about things, because you count things. And so counting is numbers of things. Like the number of ways to combine two dice rolls is a problem for counting.

One property of counting is that the numbers and the thing counted are separate. In other words the thing being counted does not matter when we are counting, as long as they are countable. I think that much is clear. Numbers work the same regardless of the thing being counted.

So let’s then define how counting works. Let’s say the cardinality of a set determines the “nth-ness” of the number, and the kinds of things the set holds inside is how we determine the thing we’re counting. Together, the type of thing the set holds + it’s cardinality is how we say the nth-ness of the thing being counted.

Remember the thing and the number are separate from each other, and that the count ability is also crucial. It’s the cardinality that determines the nth-ness of the count.

So then let’s count astronauts using our rule and determine who is the nth astronaut. Neil is first because when he landed on a moon, the set of all moon landers had a cardinality of 1. And buzz is second because when he landed on the moon the size of the set of moon landers is 2. Size of a set and cardinality are the same.

A set can also be empty. This set has a cardinality of 0.

So what was the set of moon landers before Neil? It was empty. In other words, there was nobody on the moon. So if we apply our rule we say that nobody was the zeroth person on the moon.

You might say that doesn’t make sense because nobody isn’t a person, but the problem is that’s a concern for the thing and not the number. We said they are separate things.

In this case we are only really interested in the nth-ness of the number and the kind of thing the set holds.

While nobody is not a person, the empty set itself definitely exists and it definitely has a cardinality of 0.

So the zeroth person on the moon was nobody. The zeroth mile is no mile. The zeroth century is no century. Some of the these things might make sense to you and some might not. But the sense that they have or don’t have in those case stem from how we think about the thing and less about the number.

I’ll give my final example.

An experiment starts at time t0. The zeroth second. Each second that is completed grows the size of our set of seconds. Nonetheless when the experiment began the set was empty. That was the zeroth second.

It’s not an actual second, but that doesn’t matter you can still count it. No second doesn’t exist but the empty set of second does and it can be counted. And in fact it’s really hard to explain counting at all if you don’t have a concept of zeroth.

That is why a zero-day exploit is called what it is because not one full day has passed since its existence has been revealed. Would first day also work, yes that’s fine colloquially but zeroth day is definitely not wrong is what I’m saying.

That is why we start the day at 00:00 in military time. Because what the time of a day means is the size of the set of hours, minutes, second, etc… that have passed. But the count starts at the empty set.

Here’s a very funny and confusing example: The day you are born is not your first “birth day”, because a “birth day” means anniversary of your birth. However the day you are born is the empty set from which that count begins. Birth day in this sense is an overloaded term in English but in many languages it’s literally called birth anniversary.

Anyways, that’s what I have to say. Probably much more than anyone wanted or needed but I hope it was at least clear what I think. If I’m mistaken then let me know.

The original discussion was regarding there's no such thing as a zeroth X, and what I've been trying to say this whole time is sure there is, it's the beginning. Which is why you start counting time from 0.

Interesting about patient-O though. I didn't know that.

My previous comment may have seemed snarky, but that wasn't my intention. I tried to originally write something that didn't seem sarcastic but it was just long.

The best way to explain my point was to just to agree and then list the contradictions that arise, e.g. The day starts at 12:01 since there's no zeroth minute, etc... and that unfortunately has the effect of looking like snark.

https://r3ply.com