Tailwind removes the C and the S, leaving only the S in CSS. (In long form, CSS is supposed to encourage modularity and a separation of concerns, concepts totally abandoned in the world of per-element styles.)
Let's subject that claim to some basic arithmetic.
A university would expect maybe eight undergraduate level courses, four credit hours per course, and you are expected to do three hours of study per hour of credit per week. That's 1,500 hours.
How does that compare to the "hours played" on a typical strategy game Steam review?
much less dopamine, video games are realy good at stimulating that. Studying physics not so much.
Don't get me wrong Physics can be satisfying in its own ways but the amount of dopamine rushes you get from cracking a hard problem are few and far between. Often enough you just get frustrated and once you find the solution you think to yourself oh is that easy how didn't i get that before and feel absolutely stupid.
> In practice I don't think any harness (happy to be corrected here!) uses the lesser capability models for writing code. The cost trade-offs are rarely worth it.
That's not something I understand very well. The less expensive models will quite happily chug away at tasks, if the codebase is well-structured (small files help a lot) and your instructions are clear. In contrast, I've never seen a large model turn bad instructions (instructions that would cause a human to think before starting) into a result I liked. You can run small models almost 10-100x as long for the same price in dollars, which covers a lot of correction and adjustment.
Why does everyone say the trade-offs are rarely worth it?
> In contrast, I've never seen a large model turn bad instructions (instructions that would cause a human to think before starting) into a result I liked
I think the distinction is here.
I expect my agent to build from product level descriptions. This might include specific special cases that I call out, but will rarely highlight existing special cases or edge cases - they already exist in the code, and I'd expect a programmer to make sure that behavior continues to work.
If a feature hits lots of these edge cases, the weaker model that is reading the code (aka Haiku) won't understand their significance, and will report back to the planning model incomplete or incorrect information.
The planning model (Opus - which hasn't actually seen the code remember!) will build a plan that is incorrect or incomplete and delegate coding to the mid level model (Sonnet) which will do it's best to make things work, without understanding the overall picture.
This is how you end up with slop - for example Sonnet reimplements things that already exist because it found one of the edge cases, but Opus had never known about it because Haiku didn't understand it.
It's possible that the new "agent teams" feature in Claude code can help with this. That keeps each agent alive with its context so they can ask each other things, but I haven't tried that enough to be sure - let alone with the specific model mix like this.
In your case, you are giving the Sonnet model specific instructions for what to implement mindlessly. I'd expect that to work well!
But that's not the same as the agentic workflow many other are using.
That's a great analogy - everybody knows you forget pretty much everything after a few years away from the line. It should be seen like one of those studies that proves smiling is correlated with happiness. :-)
I'm sure he's just overjoyed to be tried without representation or evidence in the court of public opinion, a big step up from being punished with no process at all.
As someone who was born, I can attest that the experience of not being born consisted of billions of years passing by in an imperceptibly short instant, followed by being five.
