geysersam

If these are such well known techniques, why didn't the qwen people already use them to fix the bug in the model? Does the weight editing deteriorate the model in some other areas?

> eliminate all 3rd party dependencies

The third party dependencies are also using AI to improve their code.

Isn't 3-SAT NP-complete?

Why do Americans accept that their democratic country is actively intervening in other countries elections? It's grotesque. The US security establishment has been allowed to run amoc for decades and they have destroyed the values that made the US country that people admired across the world.

Serbia is not the Euro-zone. It's not even in the EU.

Most of Eastern Europe and the Balkans see decent economic growth currently.

Was farming more fulfilling before machines? In some ways, most likely: During harvest the entire village gathers and works together in beautiful landscapes, using local knowledge passed down through generations.

Preferences change. If AIs can lower some of the barriers to learn and develop mathematics that might mean more people than ever will do math.

This assumes AIs will become more capable than humans in every way. If so, then yes, there's little point for people to be part of the discovery process other than to tell the AI what to investigate next.

But I think a more likely scenario is that we don't solve "general intelligence" and AIs remain these incredibly capable "tools" that can solve navier stokes but completely lack sense in other situations.

In such a scenario there is considerably more room for human creativity.

What makes you think that the number of bugs is finite? There are alternatives to that possibility. A perfect program without bugs might be like a perfect definition of the concept of "chair". Whatever definition you come up with there will likely be edges cases that makes the definition awkward or incorrect, and any definition that takes into account all those cases will not be a particularly good or useful definition.

In a security order based on mutual assured destruction, a capable missile defence system is a threat.

Where does the potassium come from?

The others are not in any way compelled to do that deep analysis and understanding. They're doing it because they find real value in the material produced.

> the inverse relationship

So you're saying if a place is more pro-palestinian it's less crap to live there. Makes sense.

This is theory. If we look at the real world, do sovereign wealth funds typically invest in cigarettes? I don't think so. I think they typically have more stringent ethical guidelines than most regular funds.

I'm typically not in favor of sovereign wealth funds anyway. Because I don't buy the idea that the free market is always more efficient than the public sector, and if that's the case then investing the funds in the public sector, schools and improving health and social outcomes, is better than buying SpaceX even if they're going to the moon.

Of course there are exceptions.. if a country has more money than they know what to do with, it makes sense to pile it.

I agree. If you used the AI and didn't bring anything of your own into it. Didn't make any adjustments, didn't follow your own ideas and vision. If you just write two sentences and waited one hour, fine, it's not your work. But most AI work today is not that. People spend hours, prompting and tweaking. It's a tool like any other.

Ridiculous. Whose work is it then? That's like saying the python script you wrote is not your work, because almost all of the hard work was done by the engineers who wrote cpython or GCC.

There were also issues recently where low river water levels forced reactors to reduce output and even shut down.

But he gave a reason for that. "Before that happens the frontier will move". Why do you think it will happen anyway? Do you think the frontier will not move fast enough that local models are unable to catch up, or do you think people will prefer local models at a point. Or something else?

They can probably afford to do both.

That sounds like a lot of work.

I long for the day when AI will just say that directly: "your soldering sucks man" instead of the bizarre made up and jargon packed language they use now.

That sounds like a marvelous scenario for economic growth, because we will have tons of capacity to deploy ai models everywhere.

The bad scenario is compute efficiently not increasing much, and hardware prices staying high, while AI is too expensive for most tasks, and 50% of all investments in the economy over the last X years have gone into AI.

Versioning seems like a separate problem, no? I mean, you can version control your python program generating SQL queries

Interesting, why do you typically prefer clickhouse local to duckdb?

Why not slap an order_by on the end of the query and you're all good? If that's expensive maybe you can add a limit clause to your source.

I've been looking at dbt for exactly this reason but don't quite get the advantages if you're not interacting with a data warehouse of some sort.

I agree sql is more elegant. The problems arise when you have to add logic on top of sql. Often I end up constructing queries via string manipulation and that is not very ergonomic. Polars api is more verbose and complex than sql but at least it's not meta-programming.

The duckdb python api is okay, but it is a bit limited, no ctes, no as of join, and it can be slow at bind/interpretation time when you do stuff like unioning multiple relations in a loop (I think that becomes O(N^2), but I might be wrong). Most issues can be worked around, but Polars is designed from the ground up to be used from python.

That was my point, replacing one random sampling by another has exactly that effect.

It has an effect, 50% of the time it's positive, increasing the quality of the answer, and 50% of the time it's negative.

It might have been a gift by his friend. You don't know the details, maybe the friend is more affluent, maybe op had already let him stay in his apartment and given him food and this was the friends covert attempt to even the scale without directly offering money.

Why would it be useless in 3 months?