bluelightning2k

That's like comparing any serverless offering to a continuously running host.

Continuously running means you're doing it wrong / it's a bad fit. But yes the ratio is somewhat extreme.

Thanks. Makes sense. Good luck with the pivot

I follow things pretty closely, and I was genuinely unsure if GPT5-pro was a real thing or just equivalent to setting GPT-5 to high reasoning.

Does the fact it's entering the API confirm that it's a fully separate thing which is not widely benchmarked, and which was not part of the GPT-5 stream?

This particular thing has not been well communicated AFAICT.

YCS19 - so 6 years. Obviously you did well enough to survive. Interesting that the pivot is something so basic after all this time. Kind of interested in the story there.

Admittedly my experience of what a battlefield comms system would look like comes only from games, but I think it needs emotes, flashy red low health indicator, and microtransactions.

I usually like his analysis but this one smacks of extreme & unnecessary bias.

Describing GPT5 as underwhelming seems subjective and somehow also wrong. It's won me and many other devs over. And Sora 2 is also clearly impressive.

UK. Remote. Senior/principal product engineer.

Built and sold £millions of my own software - including sophisticated software & infra to reliably handle millions of requests for major enterprises.

Technologies: nodejs, next, react, gemini/gpt/claude for agents, kubernetes, serverless, cloudflare stack (durable objects, workers for platforms) for alternative scaling models, etc.

Still running that company but interested in joining a senior role.

I love Cloudflare engineering writeups & the workers platform

This was very well written. Rare to see such spark in a tech write up.

I really like using Cloudflare. I think durable objects are a great innovation for example.

Spent 5 minutes trying to get basic pricing info for Moondream cloud. Seems it simply does not exist (or at least not until you've actually signed up?). There's 5,000 free requests but I need to sense-check the pricing as viable as step 0 of evaluating - long before hooking it up to an app.

This is impressive but really odd.

Isn't the entire point of Raspberry Pi to not be premium with a nice form factor, etc.

And why would I use a mechanical keyboard to drive the type of workload I'd be doing on a Pi.

Seems like they've taken super opposite and incompatible parts of PC use-cases and combined them in a really odd way.

Great industrial design. Which again isn't something I'd want from a Pi. But at the same time we all appreciate.

I kind of like it but do find it baffling.

I have never understood how Oracle remains a big company. Seems to me they have this type of reputation and just nobody will ever willingly buy anything from them ever

Some people actually accepted the contracts before the uno reverse llamabot could activate and block them

Just looked them up. Their pricing is around buying "coins" with no transparency as to what that gets. Hard pass

This is awesome. I still use windsurf and like it. Their tab model is really good

Logically it can only be API call based, as you bring your own IDE plugin. So there's no possibility it's based upon any UI level concept such as top level messages. The subscription wouldn't even necessarily know.

What a productive second you must have had

100% Java compatible

Claude Code and Gemini CLI (and OpenAI Codex) are first party from the respective companies. But also kind of products - in extreme cases people pay $200/month for Claude Code and get $thousands and thousands of usage. There's product bundling there beyond just the interface.

I think Claude Code specifically has a reputation for being a 1st class citizen - as in the model is trained and evalled on that specific toolcall syntax.

Is this the company that did shady things by buying an open source repo and kicking out the contributors? Something to do with OpenCode or SST or something idk. Could be a different company ?

Counter-take: this is a good thing.

Seems like some people are account-sharing or scripting/repackaging to such an extent that they were able to "max out" the rate limit windows.

Ultimately - this all gets priced in over time; whether that's in a subscription change or overall rate limit change, etc.

So if you want to simply use it as intended, over time stopping this kind of pattern is better for us?

"We were going to implement a counter but hit out weekly Claude code limit before we could do it. Maybe next week? Anthropic."

We have x.ai and z.ai, but why?

(Sorry... dad joke.)

This is fun - although I think the scoring system could do with some weighting.

"Dematerialize matter from one location and then rematerialize it in a second location. "

is worth the same as "Automatic sliding doors"

This is (slightly) old news from July 8, resurfaced due to the Qwen 3 coder.

I think the gist of this thread is entirely: "please do the same for Qwen 3 coder", with us all hoping for:

a) A viable alternative to Sonnet 3 b) Specifically a faster and cheaper alternative

No, guessing it just fell off the algorithm (unless for some reason YC felt it was harming Cursor or whatever)

Very interesting to see Github Copilot so close to the top here. GHCP is weird because it basically "won" out the gate and then just disappeared from mindshare so hard that nobody's even mentioned it on this thread.

That indeed is very surprising. I guess they're larger companies (even getting their own engineers to star in Google's case is a pretty big deal).

From HN conversations I'd have said Claude Code was 90%+ of the market by what I remember seeing.

Is that even true though? As per another comment Gemini CLI and OpenAI Codex have more stars, and they're both much larger companies.

Then Replit, V0, Lovable are all extremely popular.

I don't have a horse in this race just find it interesting. Your Coke vs Pepsi analogy is a good one though.