Holy AI, this shit is expensive. I was a bit suspect (no experience too) so I run some Claude calculations and it's also giving me a $350-450k to run GLM-5.2 at full precision (un-quantized). For rental on Azure, it's giving me $96–$144/hr. That translates to $22/M tokens which way more expensive than API pricing at z.ai. To get close to API pricing, you have to seek cheaper providers but that only gets you close to z.ai pricing not lower.
Caveat here is that all of this is Claude math, but would be interested in someone more knowledgeable of the math chiming in. I was thinking that API pricing was highly inflated in order to cover subscription costs but with these calculations it might be not?
That will make things worse unless you reign in crypto. If you regulate Kalshi, people will use the unregulated polymarket which is already illegal in the united states.
I made an article a day back how Polymarket is not exactly fair when it comes to market resolution (https://omarabid.com/polymarket-bet/) They actually don’t decide on market outcomes, they outsource that to UMA which is as opaque as it can get.
Consider the probabilities in polymarket as the prob. of the polymarket market itself outcomes rather than reality.
I think you have it the other way around? D1, DO, KV are lockin. The worker is not lockin as it's just JavaScript/WASM and can run in a regular browser.
To be honest their service is so cheap, that it's extremely unlikely to get such an attack without costing the attacker an equally equivalent amount.
Also, if it's really a problem or you are estimating it'll certainly be a problem, do email alerts or a strict shutdown if xx requests are hitting your account. You can have a simple counter in a KV.
I think that's when selling inside America but I don't remember seeing any american company proudly advertising its product as "American". I'd wager that today they want to hide that fact.
Easy fix: Code's basically free now, so just pipe your errors straight into an LLM and get instant patches. Sure, the patches themselves are broken too, but no worries! just pipe those back in again. Code's disposable now, fresh code generated on every request.
On a more serious note, I think the problem will be the inability to handle/maintain the systems once they are too big and nobody has no idea what's inside of them or what they do.
Yeah, it’s so easy to generate code that you can do a whole codebase rewrite in a day.
Is this a good idea? Probably not—in the past we would only do that when the architecture was causing serious problems since it always has tons of behaviors that will accidentally not get carried forward, some of which are load bearing and will cause bugs.
Now we can do it in an afternoon and get the same long term bug behavior.
The trick to talking to strangers, in my experience, is to find an opportune moment where you have a reason to talk to them (ie: you are both looking at the same departure timetable and you ask about a particular train/flight). The response will determine whether the stranger is open to carrying on a conversation. Of course, if their presence in the shared space is short, this makes it harder as these opportunities are not always present.
Smoking is probably the best lubricant (ie: borrowing a lighter, asking about a brand/vape, etc.) and people when they smoke are usually more open to strike a conversation. That's not an endorsement of smoking (and I've quit very recently).
Caveat here is that all of this is Claude math, but would be interested in someone more knowledgeable of the math chiming in. I was thinking that API pricing was highly inflated in order to cover subscription costs but with these calculations it might be not?
reply