For toy and low effort coding it works fantastic. I can smash out changes and PRs fantastically quick, and they’re mostly correct. However, certain problem domains and tough problems cause it to spin its wheels worse than a junior programmer. Especially if some of the back and forth troubleshooting goes longer than one context compaction. Then it can forget the context of what it’s tried in the past, and goes back to square one (it may know that it tried something, but it won’t know the exact details).

That was true six months ago - the latest versions are much better at memory and adherence, and my senior engineer friends are adopting LLMs quickly for all sorts of advanced development.

They’re not bad, it just depends on how much you process them. Oven baked fries for instance, are not far removed from normal oven roasted potatoes. Par boiling and then deep frying them in fats will release more nutrients than you would get from a normal baker. Slicing paper thin and perfectly frying is an industrial process that most people can’t replicate at home.

I honestly think home cooked potatoes are going to be perfectly fine in most ways.

>Slicing paper thin and perfectly frying is an industrial process that most people can’t replicate at home.

It's baffling how for some people, the only way they can explain why chips are unhealthy is "industrial process", when the explanation is pretty obvious: thin slices means more surface area, which means more oil absorption and burnt bits. If you replicated the thiness at home somehow (which isn't hard if you have a mandolin), it'll be equally as unhealthy, maybe more if you factor in that your temperature control wouldn't be as precise.

Why would industrially or restaurant made ones be any nutritionally worse? Processing is processing no matter who does it and processing does not automatically make something less healthy or raw meat would be healthier than cooked.

Maybe go try to meet some truly poor people and understand their story. It might provide you enough context for this discussion.

You are claiming sophistication is quality, it is not.

Works great for me on Frigate. I am doing object detection on 3x 4k streams and it’s only 20% utilized.

> What would change for me is how tactful I am in wording my response to it

So code is not code? You’re admitting that provenance matters in how you handle it.

I don’t think you can put pump breakage on an agenda. Sometimes emergencies require emergency action.

The pump broke in June, the meeting was in August. Does it take more than a full calendar month to update an agenda?

That’s not what the parent commenter asked though, they wanted a price for not being concerned about limits. The API pricing is that.

I doubts thats what they want. They want a static fixed price, $5k a month for example and never have to think about it.

Take the API and assume 24/7 usage (or whatever working hours are). That’s your fixed cost.

It’s more likely that this sum is higher than they want. So really it’s not about predictability.

Even if you used the API 24x7 for a single session (no parallel requests) I doubt you'd be able to hit $5k/mo in usage for Claude 4 Sonnet.

The way these work is they're net profitable given all users, so you have to recategorize users in one of two ways:

- a user subsidizing other users

- a user subsidized by other users

I don't know what OP prefers, but given that people are saying "woof, API pricing too expensive", it sounds like the latter.

The problem, of course, is the provider has to find a market where the one sustains the other. Are there enough users who would pay > $200/mo without getting their money's worth in order to subsidize users paying the same rate, but using more than the average? I think the non-existence of a higher-tier plan says there probably isn't, but I don't want to give too much credence to markets, economics, etc.

Just use the API

Just buy multiple Max subscriptions.

> SpaceX making such leaps and bounds over NASA

What? These two do two separate things, do you mean ULA?