I do regular evaluation of both codex and Claude (though not to statistical significance) and I’m of the opinion there is more in group variance on outcome performance than between them.

Chinese gold has been tungsten spiked, fairly often actually, but it’s a known fraud vector there so is broadly audited.

That is I get more variance between opus 4.6 and itself than I do between the sota models.

I don’t have the budget for statistical relevance but I’m convinced people claiming broad differences are just vibing, or there are times when agent features make a big difference.

The llms know about every previous disclosed security vulnerability class and can use that to pattern match. And they can do it against compiled and in some cases obfuscated code as easily as source.

I think the security engineers out there are terrified that the balance of power has shifted too far to the finding of closed source vulnerabilities because getting patches deployed will still take so long. Not that the llms are in some way hampered by novel code bases.

So I shouldn’t say it never happens.

I think the position that ai is morally troubling enough that the downsides out way the positives is perfectly defensible. But the entire argument becomes a joke when you can’t accurately catalog the positives.

Dont justify after the fact just dumbly implement the rule.

The problem with the post ww2 world is that the old definition of winning a war no longer holds. You just don’t see wars of conquest very often and they don’t seem to work when they happen.

The closest I can think to winning off hand is a few of the colonial civil wars. Vietnam for instance won in the sense that they outlasted the US and have a nominally communist government but it is not an outpost of the Soviet Union and it’s a major trading and tourist partner of the US.

Iraq is not led by a belligerent to the US dictator and Afghanistan isn’t home to training camps for terrorists dedicated to attacking the US (yet).

These were all extremely stupid, expensive and inhumane military actions. But the US never went into them to hold territory. So “there until we got tired of it” is as close to winning as it was ever going to be.

This is as near an iron law as there is economics, you can’t keep a peg and have a large trade in a large liquid commodity market. China is trying to slowly thread this needle and they can get away with it with Iran and Russia because they are approaching vassal status because the petrodollars are closed to them. Everyone in the world can see this and wants to avoid it.

If you are an oil producer what you want is to diversify your currency risk. Right now China is _preventing_ this, because there is no way for them to become a major player in that market without huge impact on their economy and probably their political system.

All of that happening with the worlds biggest oil producer, its second biggest manufacturer, who is food independent and has the worlds most powerful military just lets it happen. And no shooting war breaking out between them.

I’m betting on slow currency liberalization and a transition to a multi currency petroleum industry and subsequent inefficiencies in global trade. But feel free to bet how you want.

So no one is going to use a controlled currency for a hard liquid commodity. So if China wants petro yuan they have to liberalize that, which will break their peg.

China could have more international trade in the yuan before all of Americas recent misadventures. But that has cast consequences for their economy, and possibly the ruling elites power structures.

So no one is going to take up a lot of yuan trade unless that changes or they are forced to.

But that puts China in a bind. Liberalizing their currency is going to require very careful and slow actions, China threads this needle now in a very fraught way. If they openly start trading oil at any real size in yuan that will break their peg as you’ll be able to trade through the oil markets.

This is the main reason there isn’t more petro yuan already, it’s bad for China.

I do head to head comparisons with this setup pretty regularly and what I’ve found is there is not much difference in outcomes between the 2 frontier labs at equivalent model settings. It’s hard to get statistically significant results on my budget and eval ability but my anecdotal feeling is that there is as much difference in group as out in outcomes.

Given that setup I use codex much more than Claude because it’s more reliable.

But I believe it’s easier to go from nothing to decent with Claude.

For other stuff I use Claude.

The CME uses a system where orders (and fills) are entered via a direct TCP/IP connection between the trading system and the exchange. There is no opportunity for any system besides the trading systems (computers, switches, etc) on each side to see the order.

The CME then distributes the market data to a variety of paths in the form of price updates to an order book. That is the price and quantity of buy and sell orders at each price point. While these data feeds aren’t public, the cme gets paid for the data, it is widely disseminated, but it’s had the trading system identification removed. This whole story comes from people with market data contracts observing these feeds.

There are other exchanges that provide more information about who is entering orders but infrequent but large participants don’t like this because it allows the market makers an information advantage against them. And large block participants tend to correlate with real economic activity (oil producers and consumers for instance).

If a market maker knows an order is for one of these participants they can presume the size is going to be bigger and thus charge more for liquidity (in the form of of widening the price on their bid/ask spread offerings). Half the job of a modern HFT is predicting this sort of thing.

So the cme keeps it this way so one set of their market will keep using them. It’s part of the balancing act a two sided exchange has to navigate.

China is for sure providing material for drone and rocket manufacturing as well as air defense systems.

https://moderndiplomacy.eu/2026/01/28/how-china-aims-to-bloc...

“Treason against the United States, shall consist only in levying War against them, or in adhering to their Enemies, giving them Aid and Comfort.”

So no. None of those things is treason. This administration throws the word around loosely, like many authoritarian regimes, but the founders had direct experience with a dictator who used treason charges as a political weapon and tried to ensure that wouldn’t happen here.