Warned? I knew this was going to happen and want to see more of it. I don't believe in government enforced discrimination based on country of origin. If you can't compete with someone overseas that's on you, don't make it my problem and don't go crying to the government to force others to hire you.
They didn't. It's called a Schelling point to solve the coordination problem. You don't get the luxury of picking and choosing your Schelling points a la carte. They come rarely and when they come you have to act or the window passes.
The US and Iran are very different countries. You can't just fix one variable to be the same in a hypothetical and expect us to nod along as if this reveals any insight. It's a shitty rhetorical tactic.
For my longer spec files, I grep the subheaders/headers (with line numbers) and show this compact representation to the LLM's context window. I also have a file that describes what each spec files is and where it's located, and I force the LLM to read that and pull the subsections it needs. I also have one entrypoint requirements file (20k tokens) that I force it to read in full before it does anything else, every line I wrote myself. But none of this is a silver bullet.
It's easy to know why they work. The magic invocation increases test-time compute (easy to verify yourself - try!). And an increase in test-time compute is demonstrated to increase answer correctness (see any benchmark).
It might surprise you to know that the only different between GPT 5.2-low and GPT 5.2-xhigh is one of these magic invocations. But that's not supposed to be public knowledge.
They never parsed your prompt. The magic word reduces the probability that the token corresponding to the end of chain-of-thought will be emitted, which increases test-time compute.
Blue light blockers are a scam that was created when some circadian rhythm research went viral (in a highly misrepresented way) online a few years ago. It's a stunt to make some quick cash from unwitting buyers.
To the extent that there's some sort of clinical statistically significant lift over, say, however you might control for a placebo here, who knows? To the extent that it's working for you in a meaningful way, placebo or not... does it even matter? If they're working for you, they're tautologically not a scam, except insofar as you find yourself missing out on benefits promised that aren't being realized, or you feel that the price you've paid is disproportionate to the benefit because some much cheaper option exists in some sense.
Put another way, placebo or not, if there's an effect, and it's a positive one for you, it really doesn't matter. It's working.
That won't work for immediate relief; placebo is studied statistically as a cumulative change over a period of time. Not when I put Gunnars over my sore eyes from computing too much and get immediate relief. Frankly, I find this discussion a bit insane, a bunch of people trying to persuade me "it's all in my head" because they align with the opposite opinion, not with reality.
This is the first semblance of policy certainty. The ruling is a good thing for everyone, Republicans and Trump included, even if they're not intelligent enough to understand why.
reply