They still don’t support json schema or batch api. It’s like deepseek does not want to make money

What do you currently use for json and batch, I was doing some analysis and my results show that gpt-oss-120b (non batch via openrotuer) is the best for now for my use case, better than gemini-flash models (batch on google). How is your experience?

I don’t get why people are so surprised. Didn’t they learn anything from Twitter APIs and the like. The APIs are open as long as they serve the short term problem then Anthropic builds the features people actually use (more or less) and ban the usage of APIs for competing clients

This is the exact opposite of what you are describing, where Anthropic locked down the API until they released their competitor, then re-allowed API

Not really. They realised they acted too soon. Give them some time until the market “consolidates” and they will change again their policy. Why would they want someone else to develop competing clients?

Codex is just better than Claude but Claude is faster and has the better UI for vscode. That’s why I use Claude as main coder with codex(5.4 with xhigh effort) as mcp reviewer etc. It is clear to me that codex is a better programmer but the UI and speed are too much of a con to use it exclusively. Claude is just clumsy

The world would barely notice if both OpenAI and Anthropic “fail”. There is a lot of competition in this space

what makes gpt 5.4 bad to chat ? To me it seems smart and does the job albeit it’s a bit slow. I’m using it only/mostly with the “pro” /xhigh reasoning

In my experience, within one reply ChatGPT (compared to Claude) tends to:

* generate a lot of text

* answer at least a few of what it thinks might be follow up questions

* restate its original answer a few times

* suggest a follow up “if you want, I can turn that into…”

It feels very tedious and noisy.

Yeah, this perfectly sums it up.

Especially the “If you want” at the end of every single reply.

The way it converses is the least human like out of all models. It communicates like its writing markdown documents instead of just conversing normally like every other model does. You ask a question and it spits out a design doc instead of just answering the question like a normal human would.

But why would you want it to converse like a human? What I need is concise information not a buddy to talk to. If anything I see it as a feature

I think it’s the default system prompts of ChatGPT. I do think it’s better with the “Professional” tone with “Less Emoji” but that’s just me.

It’s not always about money. It’s also about the time of the developer. Even for a hobby project you may burn out before to actually deliver it.

I'll say depends. Personally, my hobby projects are about me, just shared with the world because I believe in Free Software.

Yet, I'm not obliged to deliver anything to anyone. I'll develop the tool up to the point of my own needs and standards. I'm not on a time budget, I don't care.

Yes, I personally try to reach to the level of best ones out there, but I don't have a time budget. It's a best effort thing.

In reality you are always on a time budget that is correlated with the output of the software you develop.(I.e is it worth it your time?) I’ve found out that the most important thing is to get feedback early even from yourself using whatever software you develop. If you develop a small effort piece of software you can ship it before other stuff is starting to compete for your time. But if it takes a year or more before even you can make any use of it I guarantee you that the chances of shipping it diminishes significantly. Other stuff competes for your time(I.e family, other hobbies etc).

I think we tackle the same problem in different ways. For me, if something is not urgent, I do it in a best effort way, and the shipping time doesn't matter.

I generally judge whether I allocate time for something or not depending on the utility and general longevity of the tool. I hack high utility / short life tools, but give proper effort to long life tools I need. As a side-effect, a long life tool can start very crude and can develop over time to something more polished, making its development time pretty elastic and effort almost negligible on the long run.

For me shipping time is both very long (I tend to take notes and design a tool before writing it), yet almost instant: when I decide that the design is enough for V1, I just pull my template and fill in the blanks, getting a MVP for myself. Then I can add missing features one at a time, and polish the code step by step.

Currently I'm contemplating another tool which is simple in idea, but a bit messy in execution (low level / system programming is always like that), but when it's design is over, the only thing I'll do it is to implement it piece by piece, without no time crunch, because I know it'll be long-living tool.

I can time-share my other hobbies, but I have a few of them. I do this for fun. No need to torture myself. And, I can't realize my all ideas. Some doesn't make sense, some doesn't worth it, some will be eclipsed by other things.

That's life, that's fine.

Times are monies though

I think dropping a nuke is not out of question. We just have to watch the language of the U.S administration over the next weeks.

It's very hard to imagine nuclear weapons being used in this conflict.

It was hard to imaigne the United States sticking their nose in the middle east again

It was hard to imagine America kidnapping a head of state, the president of Venezuela

It was very hard to imagine Amercia threatening annexation of Canada and Greenland

Your difficult to imagine world is closer to reality than I think either of us would like

Agree with the general point, but the second example shouldn’t have been that hard to imagine. We’ve done it before: https://en.wikipedia.org/wiki/United_States_invasion_of_Pana...

And is true about the nukes

Is it? Israel is a rogue state at this point, and it has openly stated that were the US to withdraw, they would use "any means necessary".

They seem to not be satisfied by causing global economic harm, they want nuclear annihilation as well.

If they do, they will be exterminated.

There is no timeline where Israel comes out victorious if they start using nuclear weapons.

What quote are you referring to?

yeah, I think this will escalate further and US might drop a tactical one on some island (military but not heavily inhabited) as a show of power. I don't see a way back out otherwise for US.

The US doesn't need to show power. Using nukes is all downside.

Normally yes, but this is the Suez canal moment for US. I think countries stopped caring about US especially in the last year and starting reducing dependence on US economically and politically. So, the unpredictable Trump may try to do something dramatic. China will be the biggest winner of this all.

The world sentiment is to deintegrate with the US - something Trump campaigned on - and so he is going to do something dramatic? And that dramatic thing is to use nuclear weapons?

Just no. Stop this.

What would that gain?

I think they will settle that Iran won’t reuse whatever they have of that nuclear fuel if the U.S drops the nukes on the Iranian nuclear sites. Iran may also think twice before rejecting a deal in which they agree to not acquire a nuclear weapon.

>> Grounding identity in a phone number is very reasonable for almost all normal usage.

Yeah if you compare that with Facebook messenger and other such services but if you want secure communication it's not reasonable.

I don't think they would be so much more expensive but they would be less profitable for sure and perhaps less "innovative" as a big chunk of the profit will go into regulation stuff.

As a Go developer I would use this unless...I would have to build the std libraries myself. Now I also wonder how this compares with tinyGo. Why would you pick one over the other...