Yeah, you're right, and the snark might be warranted. I should consider it the same as my stupid (but cute) robot vacuum cleaner that goes at random directions but gets the job done.
The thing that differentiates LLM's from my stupid but cute vacuum cleaner, is that the (at least OpenAI's) AI model is cocksure and wrong, which is infinitely more infuriating than being a bit clueless and wrong.
I've been trying to solve this by wrapping the generation in a LangGraph loop. The hope was that an agent could catch the errors, but it seems to just compound the problem. You end up paying for ten API calls where the model confidently doubles down on the mistake, which gets expensive very quickly for no real gain.
You can play with the model for free in chat... but if $20 for a coding agent isn't effectively free for use case it might not be the right tool for you.
ETA: I've probably gotten 10k worth of junior dev time out of it this month.
It was there a few (<5? I think?) minutes after the Anthropic post went out. If you look at Windsurf's web traffic it looks like they did a thing (model is an int) to make it so the IDE doesn't need to update to get new models.
The price shown is dynamically updated to today’s dollars, and the subscript is the original value at the publishing date of the source text.
> inflation adjustment: Inflation.hs provides a Pandoc Markdown plugin which allows automatic inflation adjusting of dollar amounts, presenting the nominal amount & a current real amount, with a syntax like [$5]($1980).
I don’t understand this article! PC motherboards with 10GbE ports have existed for years in premium offerings? Is this notably cheaper than the current chip they use?
But $10 is not a cheap price point - if it's a component on a motherboard, it really needs to be sub 25 cents in 10k volume orders before motherboard manufacturers start shoving it into mid level boards just to have one more bullet point on the spec sheet of a motherboard which sells for $50.
The current 1G components are already 4x-8x that cost in volume. Much like 100M never seemed to quite go away for decades (and still hasn't in certain areas), this doesn't need to hit the bottom barrel to still hit cheap motherboards. Especially for people buying by components, many "entry level" A620 motherboards launched 2 years ago already had 2.5G NICs.
It may be a bit longer before random PCs at Walmart have 10G more often than not but it won't be long at all for "mid range" motherboards you're talking about.
Most of my flights (within the USA) are on ERJ-175. I even switched airlines midway through the year and this was still true. I only see Airbus and Boeing for transcontinental flights.
Posted up above, here's a collection of English pronunciation rules that English speakers have internalized so well they can't generally explain them: https://www.zompist.com/spell.html
"Ghoti" is mentioned a few times there, but basically "fish" is a nonsensical pronunciation that breaks several rules. There's a reason (well, a few reasons) why if you ask English speakers how to pronounce "ghoti" and they've never seen it before, they'll probably all guess some variation of "go-tee" or "go-tie".
That's such a dumb example because it claims to follow english rules for those letters while ignoring the actual rules. It makes a somewhat humorous joke, but people pretending that it means anything linguistically are either ignorant or intentionally trying to confuse people.
The concern with Esc is that if you hit more than 3 times the user will be stuck on the page. The first 3 presses would trigger the redirect, the 4th press would be intercepted by the browser and stop the page load.
It sounds like you know what the problem with your AI workflow is? Have you tried using an agent? (sorry somewhat snarky but… come on)
reply