Hacker News Reader

Comment by Chyzwar | Hacker News Reader

Anthropic is building moat around theirs models with claude code, Agent SDK, containers, programmatic tool use, tool search, skills and more. Once you fully integrate you will not switch. Also being capital intensive is a form of moat.

I think we will end up with market similar to cloud computing. Few big players with great margins creating cartel.

View on HN

>Anthropic is building moat around theirs models with claude code, Agent SDK, containers, programmatic tool use, tool search, skills and more.

I think this is something the other big players could replicate rapidly, even simulating the exact UI, interactions, importing/exporting existing items, etc. that people are used to with claude products. I don't think this is that big of a moat in the long run. Other big players just seem to be carving up the landscape and see where they can can fit in for now, but once resource rich eyes focus on them, Anthropic's "moat" will disappear.

I thought that, too, but lately I've been using OpenCode with Claude Opus, rather than Claude Code, and have been loving it.

OpenCode has LSPs out of the box (coming to Claude Code, but not there yet), has a more extensive UI (e.g. sidebar showing pending todos), allows me to switch models mid-chat, has a desktop app (Electron-type wrapper, sure, but nevertheless, desktop; and it syncs with the TUI/web versions so you can use both at the same time), and so on.

So far I like it better, so for me that moat isn't that. The technical moat is still the superiority of the model, and others are bound to catch up there. Gemini 3 Preview is already doing better at some tasks (but frequently goes insane, sadly).

> coming to Claude Code, but not there yet

Wasn't this released a couple of weeks ago?

It was added in 2.0.74 according to the changelog, but it's not functional: https://news.ycombinator.com/item?id=46357036. I've certainly not been able to use it.

Good call. Looks like there's at least a user submitted workaround now [1]

[1] https://github.com/anthropics/claude-code/issues/14803#issue...

Also: Claude has already asked me several times the last few days if I want to install an LSP for various things. I've not seen any signs of LSP use yet.

Does OpenCode work with their subscription Max plan or is it API pricing only?

Pretty sure it does. It just uses whatever you log in as.

It can, but the auth & communication to Anthropic's APIs is basically reverse engineered from Claude Code. While it works, and it seems Antropic is choosing to look the other way, it _may_ result in your account getting banned, as I'm pretty sure it's against their TOS.

I haven't experienced this myself, but RooCode does something similar to OpenCode's approach and the maintainer has reported some bans [1].

Google on the other hand, is being very strict about keeping you locked in to their tools, unless you use API keys, of course.

[1] https://github.com/RooCodeInc/Roo-Code/pull/10077#issuecomme...

LibreOffice didn't replace MS Office, and Octave didn't replace Matlab. It seems to me that there is even less of a moat with these products than there is with Claude Code, yet neither was commoditized.

Google Workspace replaced Microsoft Office. It has around 70% market share. Microsoft Office is still dominant in much of the traditional enterprise, but the moat is shrinking.

I can use Claude in Jetbrains IntelliJ and in Zed, I can use it with OpenCode, and there are lots of other agent tools. Everyone can build these tools around an LLM, and they're already being commodified.

The moat right now is the quality of the model, not the client. Opus is just so much better than the competitors, at least for now.

Enterprise is what determines commoditization as that's where the lion's share of the revenue comes from. Maybe it will eventually happen to MS Office, maybe it won't, but until it happens it hasn't happened. Access to the full MS ecosystem, technical support and seamless integration, these things matter a lot to businesses, and I'm not even saying you're wrong, but I'm not convinced yet that something similar won't play out with coding agents.

A GPT wrapper isn't a moat.

Most things are wrappers around RDBMSs.

Most true and interesting comment I've read on HN in a while!

A generic wrapper is not a moat, but the context is. Both the LLM provider and the wrapper provider depend on local context for task activities. The value flows to the context, the LLMs and wrappers are commodities. Who sets the prompts stands to benefit, not who serves AI services.

If AI is capable of doing what they claim then these aren‘t moats because they are just one prompt away from being replicated.

All those features are basically prompts in various formats, not much of a moat.

Except most of their product line is oriented towards software development which has historically been dominated by free software. I don't see developers moving away from this tendency and IMO Anthropic will find themselves in a similar position to JetBrains soon enough (profitable, but niche)... assuming things pan out as you describe.

Reply on news.ycombinator.com