[CuriouslyC]

Anthropic is going to be on the losing side with this. Models are too fungible, it's really about vibes, and Claude Code is far too fat and opinionated. Ironically, they're holding back innovation, and it's burning the loyalty the model team is earning.

[empath75]

I think you have it exactly backwards, and that "owning the stack" is going to be important. Yes the harness is important, yes the model is important, but developing the harness and model together is going to pay huge dividends.

[athrowaway3z]

https://mariozechner.at/posts/2025-11-30-pi-coding-agent/

This coding agent is minimal, and it completely changed how I used models and Claude's cli now feels like extremely slow bloat.

I'd not be surprised if you're right in that this is companies / management will prefer to "pay for a complete package" approach for a long while, but power-users should not care for the model providers.

I have like 100 lines of code to get me a tmux controls & semaphore_wait extension in the pi harness. That gave me a better orchestration scheme a month ago when I adopted it, than Claude has right now.

As far as I can tell, the more you try to train your model on your harness, the worse they get. Bitter lesson #2932.

[philipwhiuk]

> I'd not be surprised if you're right in that this is companies / management will prefer to "pay for a complete package" approach for a long while

I mean I suspect for corporate usage Microsoft already has this wrapped up with Microsoft & GitHub Co-Pilots.

[AJ007]

OpenAI, Anthropic, Google, Microsoft certainly desire path dependence but the very nature of LLMs and intelligence itself might make that hard unless they can develop models which truly are differentiated (and better) from the rest. The Chinese open source models catching up make me suspect that won't happen. The models will just be a commodity. There is a countdown clock for when we can get Opus 4.6+ level models and its measured in months.

The reason these LLM tools being good is they can "just do stuff." Anthropic bans third party subscription auth? I'll just have my other tool use Claude Code in tmux. If third party agents can be banned from doing stuff (some advanced always on spyware or whatever), then a large chunk of the promise of AI is dead.

Amp just announced today they are dumping IDE integration. Models seem to run better on bare-bones software like Pi, and you can add or remove stuff on the fly because the whole things open source. The software writes itself. Is Microsoft just trying to cram a whole new paradigm in to an old package? Kind of like a computer printer. It will be a big business, but it isn't the future.

At scale, the end provider ultimately has to serve the inference -- they need the hardware, data centers & the electricity to power those data centers. Someone like Microsoft can also provide a SLA and price such appropriately. I'll avoid a $200/month customer acquisition cost rant, but one user, running a bunch of sub agents, can spend a ton of money. If you don't own a business or funding source, the way state of the art LLMs are being used today is totally uneconomical (easy $200+ an hour at API prices.)

36+ months out, if they overbuild the data centers and the revenue doesn't come in like OpenAI & Anthropic are forecasting, there will be a glut of hardware. If that's the case I'd expect local model usage will scale up too and it will get more difficult for enterprise providers.

(Nothing is certain but some things have become a bit more obvious than they were 6 months ago.)

[AJ007]

Thinking about this a little more -> "nature of LLMs and intelligence"

Bloated apps are a material disadvantage. If I'm in a competitive industry that slow down alone can mean failure. The only thing Claude Code has going for it now is the loss making $200 month subsidy. Is there any conceivable GUI overlay that Anthropic or OpenAI can add to make their software better than the current terminal apps? Sure, for certain edge cases, but then why isn't the user building those themselves? 24 months ago we could have said that's too hard, but that isn't the case in 2026.

Microsoft added all of this stuff in to Windows, and it's a 5 alarm fire. Stuff that used to be usable is a mess and really slow. Running linux with Claude Code, Codex, or Pi is clearly superior to having a Windows device with neither (if it wasn't possible to run these in Windows; just a hypothetical.)

From the business/enterprise perspective - there is no single most important thing, but having an environment that is reliable and predictable is high up there. Monday morning, an the Anthropic API endpoint is down, uh oh! In the longer term, businesses will really want to control both the model and the software that interfaces with it.

If the end game is just the same as talking to the Star Trek computer, and competitors are narrowing gaps rather than widening them (e.g. Anthropic and OpenAI releases models minutes from each other now, Chinese frontier models getting closer in capability not further), then it is really hard to see how either company achieves a vertical lock down.

We could actually move down the stack, and then the real problem for OpenAI and Anthropic is nVidia. 2030, the data center expansion is bust, nVidia starts selling all of these cards to consumers directly and has a huge financial incentive to make sure the performant local models exist. Everyone in the semiconductor supply chain below nvidia only cares about keeping sales going, so it stops with them.

Maybe nvidia is the real winner?

Also is it just me or does it now feel like hn comments are just talking to a future LLM?

[CuriouslyC]

That was true more mid last year, but now we have a fairly standard flow and set of core tools, as well as better general tool calling support. The reality is that in most cases harnesses with fewer tools and smaller system prompts outperform.

The advances in the Claude Code harness have been more around workflow automation rather than capability improvements, and truthfully workflows are very user-dependent, so an opinionated harness is only ever going to be "right" for a narrow segment of users, and it's going to annoy a lot of others. This is happening now, but the sub subsidy washes out a lot of the discontent.

[quikoa]

If Claude Code is so much better why not make users pay to use it instead of forcing it on subscribers?

[cyanydeez]

If these LLMs and tools create real valuable products, where are they?

Shouldnt there be dedicated youtubers showibg us thwir skillz?

[popcorncowboy]

You're right, because owning the stack means better options for making tons of money. Owning the stack is demonstrably not required for good agents, there are several excellent (frankly way better than ol' Claude Code) harnesses in the wild (which is in part why so many people are so annoyed by Anthropic about this move - being forced back onto their shitty cli tool).

[keeganpoppen]

the fat and opinionated has always been true for them (especially compared to openai), and to all appearances remains a feature rather than a bug. i can’t say the approach makes my heart sing, personally, but it absolutely has augured tremendous success among thought workers / the intelligensia

[tom_m]

I thought Anthropic would fall after OpenAI, but they just might be racing to the bottom faster here.

[wolvoleo]

I think they're doing a great job on the coding front though

[ruszki]

I think there is a huge gap between people who has a good CLAUDE.md (or similar), or those who doesn’t.

When I first tried, the created code was garbage. Now that I slowly built my memory, several thousands of manually written examples and guidance, it can generate quite reliably, when it doesn’t need literally anything outside of those…

That being said, most of the vibe coded codebases (in reality every single one which I saw) use garbage memory, and consequently have garbage output.

So the same thing is terrible and great at the same time. People who give time, and people who is fine producing garbage (huge majority) says it’s great. People who just tried it out, and don’t have the luxury to potentially waste days and weeks, say that it’s bad. All of these are true at once.

[burnte]

Maybe for coding but the number of normie users flooding to Claude over OAI is huge.

[mikestorrent]

I think their branding is cementing in place for a lot of people, and the lived experience of people trying a lot of models often ends up with a simple preference for Claude, likely using a lot of the same mental heuristics as how we choose which coworkers we enjoy working with. If they can keep that position, they will have it made.

[tracker1]

I'm a very experienced developer with a lot of diverse knowledge and experience in both technical and domain knowledge. I've only tried a handful of AI coding agents/models... I found most of them ranging from somewhat annoying to really annoying. Claude+Opus (4.5 when I started) is the first one I've used where I found it more useful than annoying to use.

I think Github Co-Pilot is most annoying from what I've tried... it's great for finishing off a task that's half done where the structure is laid out, as long as you put blinders keeping it focused on it. OpenAI and Google's options seem to get things mostly right, but do some really goofy wrong things from my own experiences.

They all seem to have trouble using state of the art and current libraries by default, even when you explicitly request them.

[mikestorrent]

Well, let's be thankful that there's still some work to do so we can keep our jobs for a few more years!

[ThunderSizzle]

Github Copilot let's you pick the underlying model, including Claude models.

Is the problem you observed true regardless of the model you picked?

[tracker1]

I've only used the default selection, whatever it is in VS Code. Even paid for a year at one point as I was first using it with some SQL schema generation and it was pretty useful, kind of as a super auto-complete.

If the default option isn't at least arguably the best option I can't really speak to that. I would suggest that maybe metrics on a given set of technologies be done and that based on the project in use, that it should choose the best option dynamically by default. Such as C#+MS-SQL vs Node+Postgres vs Python+Matlab+DuckDB.

[eshaham78]

The competition angle is interesting - we're already seeing models like Step-3.5-Flash advertise compatibility with Claude Code's harness as a feature. If Anthropic's restrictions push developers toward more open alternatives, they might inadvertently accelerate competitor adoption. The real question is whether the subscription model economics can sustain the development costs long-term while competitors offer more flexible terms.