Ah I see, then indeed we disagree about the point of a library. You see it as storage, I see it as a place for object, I see it as a place for people to meet, with the "excuse" o objects or via objects (e.g. books).
Ack on the guardrail analogy.
On fireworks - do you just run serverless or do you take their dedicated instance. My question is how do you ensure that the model is not being manipulated to getting stuff it shouldn’t be.
but to get to the city we had to take it was a 2 hour long drive through twisted roads that made me carsick. I lived in a small town far away from the city. By the time I got there my breakfast I had quickly eaten gave me a stomach ache, I had woken up far earlier than usual and not gotten my 7 hours too.
Another factor: if you wanted to pay the fee you could just take the test over and over again until you got a great score. So kids with poor parents obviously had a huge disadvantage. Also kids who had the time and money could study for it with prep books - I did, while some of my friends were flipping burgers while still in highschool. Its not surprising I got a higher score than them, but it said nothing about my intellegence or understanding compaired to theirs.
Wikipedia says 'usually'. And, yes, you are right that the individual HDB flats are typically owned by individual families (though the government also owns some and rents them out, but that's the exception).
However, HDB still does a lot of the managing. Eg HDB manages the lifts and other common spaces. They organise getting the exterior repainted, or general upgrading works to plumbing etc.
HDB also subsidises the price of a flat when you buy directly from them. There's a reason they have maximum income caps for buying.
My unpopular opinion: HDB should declare victory and go home. They more than solved the original acute housing shortage they were created for.
Well, sure, but first semester performance is also a good predictor of second semester performance. And second semester of third, and so on.
But more to the point: If you do poorly in your first semester and drop out, then it doesn’t really matter if the SAT would have done a good job of predicting your second semester performance.
A cash advance is taking physical cash out with your credit card at an ATM or bank teller. Making an online purchase (which this is) is therefore not a cash advance.
This is not an issue specific to Linux, Windows installations can be downgraded as well. As for the mechanism, UEFI devices, alongside key updates, also get revocation list updates.
yep. claude keeps "habitually" trying to use `rg -rn` instead of `rg -n` because it was instructed to use "rg" instead of "grep" by Anthropic, but uses arguments for grep: `grep -rn`. My instructions and "memory" are not helping. "Oh, I did it again, and you've instructed me not to". Older tools are better for current "agents".
Why are they being told anything outside of the test? What is that for? Isn't “can you find this bug in this file?” also a test? It sounds like there are two kinds of tests? I'm clearly confused, I realize.
i'm tired of this narrative about switching to open source models. If people wanted to switch... well they would have switched!
The amount of effort Anthropic and OpenAI are putting behind inference, harness, building great agentic applications on top of the frontier model... IS NOT A SMALL THING.
Think about it, if you want to use an open source model to run an agent over long horizons, you need someone to offer it reliably. And if all compute goes to frontier models there is very little left for even startups to build agentic companies on top of these open source models.
Hard disagree. Opus reports to me like a student. Fable reported to me like a colleague (researcher). It genuinely seemed to pick up on nuance that the other models just don't, even when I tell them explicitly. It's been really frustrating that neither Codex nor Opus can make targetted edits to Fable's code without screwing something subtle up. For context, this is for computational geometry work, so your mileage may vary.
The WSJ ran an earlier article stating that unless you have access to a lot of data, computing power, and statisticians, you're going to lose on those bets.
Maybe. That narrower focus definitely makes your remark more defensible.
On a tangent: along with most economist, I generally prefer giving poor people money that they can spend on goods and services as they please; instead of having some pencil pusher decide what they need and giving that to them in-kind.
Steam take 30% of the purchases made via the Steam Store.
If you sell a game on Steam, you can redeem as many Steam Keys for your game as you wish.
Those keys are sold at 100% profit to you, Steam dont take any.
>Unlike most other nuclear reactors, Candu reactors don't require enriched uranium. Ottawa says Western allies are turning away from Russia, one of the world's key suppliers of enriched uranium.
Even if Canada winds up relying more on CANDU reactors than SMR's, there is a case to be made for enriching domestically. There are a lot of potential customers looking for a reliable, ethical supplier. Canada has the raw minerals, political stability, and a long record of refusing to weaponize despite having the capability.
Spain has 3x the emissions intensity of France. The Nordics (some of them) have energy that is cheap and clean like France. That's because they have base load that doesn't emit CO2 like France.
> and that got me thinking that there's potentially a different type of gambling app that ignores the money and is more of a social/prediction-making platform
We have that in the Netherlands and we do it massively for every World Cup and Euro. The app my friend groups are on, Scorito, has 1.5 million competitors for the current WC, in the country of 18 million, and that's not the only app.
Normally a group would make a small buy in, like €5, where the winner takes all, but this year the legal department at work forbade that, se that sort of gambling Is actual illegal.
The problem is, how do you know if someone is an expert in the first place? A LOT of podcasts are done with random people who claim to be experts, spreading misinformation. How does your product filter those out?
