In terms of raw token cost, I've seen a couple providers at (all prices in terms of Mtok) $0.95 input/$0.15 cache input/$5 output vs $3 input/$15 output for sonnet.
Task prices of courses will be more interesting - a dumber model may use more tokens to get to the same goal.
copy/paste doesn't tell you much - here's the text/html content they put on your clipboard if you're curious. Apparently GDocs supports this out of the box, just hides it from the selection box. Which makes sense given that it doesn't support any font.
That's exactly where my mind went as soon as I read the title. HN rules say to "use the original title, unless it is misleading". I think the original title meets the misleading bar but I can't speak for other readers.
"Private credit" is a finance term of art. It could be misleading if you don't have context for the correct definition, but that's true of many posts on this site.
It converts back to paid automatically if you had an existing paid subscription before. No other cases. In any case, this is still a valuable service they are providing for 6mo for free, which many will appreciate even if the goal is to recruit more users.
I'll second this. I used opencode + opus 4.6 + ghidra to reverse engineer a seedkey generation algorithm[1] from v850 assembly. I gave it the binary, the known address for the generation function, and a set of known inputs/outputs, and it was able to crack it.
None of it makes much sense. The model labelled as fastest has much higher latency. The one labelled as cheapest costs something, whereas the other one appears to be free (price is blank). Context on that one is blank and also unclear.
It's generally very helpful - someone else mentioned here the fundamental problem is lack of a tight feedback loop. It doesn't perfectly replicate the GH environment, but for my use case that doesn't matter and it's super nice to have.
Coke used to be mixed, bottled, and shipped out in an extremely quick timeframe. Inventory turned over fast.
I suspect the separated components wind up being equal to what a stale soda has, one that has been on the shelf. It’s like buying a soda whose sugar component has already gone stale.
Sure, the rest of the flavors are there and still fresh, unaffected by the carbonated water, but the sweetness one is off.
A couple years back I was looking for this sort of solution and ended up paying money to buy FilterBox which I've found to be good.
There are certain apps that I would love to be able to uninstall but have to keep for one reason or another, so I really appreciate apps like these which prevent attention-stealing notifications from making it through :)
It's a shame that AI is ruining certain phrases, the "You’re absolutely right" was appropriate but I've been trained reading so many AI responses to roll my eyes at that.
Task prices of courses will be more interesting - a dumber model may use more tokens to get to the same goal.
reply