I feel like we are seeing agentic coding tools morph into walled gardens with these acquisitions. Anthropic has restricted claude code usage while OpenAI has sort of let Codex fill the void. I am curious to see how this continues to evolve.
This is the whole point and the reason for the lofty valuations. Get everyone to shift their work to be dependent on these tooling, to the point they can't imagine working in any other way, and then raise prices. Tale as old as enterprise software.
Tale as old as the word "startup" even. Uber/Lyft did it with taxis. DoorDash did it with food delivery. You run at a loss for years while destroying your legacy competition by just outlasting them, then once you have cornered the market you squeeze.
I understand the cynism but it’s not the case here. Stainless isn't a case of blitzscaling or running a loss for years to destroy the competition. The motto of the company is polished and robust and we invested a lot into generating what we think are the highest quality SDKs available. We could have shipped things way, way faster if the focus on design and quality wasn’t such an essential part of the development process
No but Anthropic and OpenAI are very much trying to use their positions to destroy everyone's ability to do things without their product, make AI essential, and then jack prices. Thats the only way this becomes profitable.
This comment reminds me of the general malaise affecting the public discourse in the US these days. "Don't listen to the experts, all the context and knowledge they have makes them biased!!"
Of course Sam is "qualified" to speak to the ethos of Stainless, he was there longer than anyone except the founder.
Obviously not, I do feel more than qualified given that I actively contributed to developing the company. None of what I wrote in the comment you responded to is a secret or is based on my personal judgement, we’ve been open about our values
Skill issue. Taxi companies aren't able to innovate and adapt and improve, despite the competition from Uber, preferring instead to use lobbying and regulations too survive in a post-Uber world.
Actually, it is a marketing issue. Taxis did innovate and did improve and imo are a better product than uber today. They have an app that is no different than what you expect with rideshare apps. Actually it is better, I can schedule a ride and get a flat rate with tip already baked in to places like the airport. No need to fret about surge prices at all, what I see when I schedule it today is what I pay when it comes tomorrow or next week or next month, whenever I've scheduled it.
But, no one uses it, because uber and lyft have become kleenex or coca cola: the brand name associated with the basic phenomenon, such that consumers cannot even think about the phenomenon without thinking first of the brand and probably resorting to the brand.
Last time I was in Vegas (in 2026), there was a $40/$60 surcharge to my destination because the taxicab commission was able to enact regulations that says taxis can charge that, while Lyft/Ubers to the same location did not have to pay it. I was waiting at a smaller hotel for 20 minutes for a taxi to arrive to the taxi stand. There was a line of people an no taxis in sight. The time before that, I got hit with the ole "credit card machine doesn't work" bit.
If they wannted to innovate, every taxi stand would have a QR code to download their app, you set your destination while you're waiting in the line (if there is one), you get in the cab, the driver scans the QR code from your phone, payment happens via CC in the app, no surprises either. The phone app would show you the route, and then give an estimated price, and if there's an overage due to traffic/other problem, give me a notification on my phone. It would also let me set preferences for conversation/temperature.
I'm reading "enshitification", and it describes this cycle of first losing money but acquiring customers, then switching focus to catering to businesses, then to themselves and at that point the tool is not what it was supposedly intended to be.
This is the same startup culture. The only innovation here is finding new way to swindle customers and businesses out of money.
Exactly. The goal of any VC by definition is to return a positive return on investment. I guess you might have a handful of exceptions, funds that are environmentally conscious, but profit remains paramount.
The new owner's plan is...to sunset the paid product immediately and give customers access to tooling to be able to continue generating SDKs on their own. From Stainless's post:
As we focus on Claude Platform capabilities and connecting agents to APIs, we’ll be winding down all hosted Stainless products, including our SDK generator. Starting today, new signups, projects, and SDKs will not be available.
If you’re a Stainless customer, visit app.stainless.com/transition for help transitioning from Stainless-managed products to other options. As always, you own the SDKs you’ve generated to date, and have full rights to modify and extend them however you wish.
As a customer, all-in-all, we were pretty pleased with the outcome. Stainless was a great partner to us, even in "the end," and I'm really happy for the team.
With respect, you were manipulated (either by founders or by investors). Startups leverage employees' pro-social leanings to make them feel good about a fundamentally anti-social enterprise.
Why wouldn't getting more customers the plan? Anthropic doesn't acquire companies to have a lower market share. There is clearly a consolidation and a rush to get as much of the developer market as possible.
Claude is just a tool. My team members are each free to choose the text editor or IDE that they are happiest with. In the near future, I hope to be able to say the same for coding agents. I really like Claude, but I don't track Claude resources in our repos. If something better comes along, I'm betting it will be perfectly happy to parse the markdown of my existing memory files, and nothing in the repo itself will force anyone else to know that I switched.
It kind of blows my mind that the majority of Claude users have just accepted that CLAUDE.md is a tracked file that the whole team has to standardize on and share. Coding agents are the ultimate API. They conform to however you prefer to interact. Is anyone really expecting to enforce standard operating procedures with this non-deterministic black box of magic?
Frontier AI labs is pivoting to something that can justified their IPO. just like OpenAI shut down other services and pivot more into coding. They want to show profitability before their mega IPO.
I don’t really see where the “walled garden” complaint is coming from. Anthropic spends a lot of effort to keep you from churning through trillions of tokens on their flat-rate subscription plan, but that’s a billing detail, and one that I honestly don’t share the outrage about. The technology part of CC is still totally open: skills, MCP, etc. are all open informal standards and there hasn’t been any movement to lock that down.
This is missing the forest for the trees and it comes across as adulation for a corporation, when in fact, customers come first. Why are you trying to police people's opinion about a product?
I never said I didn't _understand_ why Anthropic is doing this. What I'm saying is nobody wants Claude Code, the product. They want access to the model. Feeling strapped to a harness you don't enjoy working with and which may limit your productivity is not a good experience for users. Voicing that dissatisfaction is perfectly valid. No, it's not "egregious" or "wrong", but it's also _something_ rather than "not anything".
Do you know what also give users access to the model without forcing lock-in? A better product. Building it is a choice Anthropic decided not to make.
Probably 99.9% of people subscribe to Max for Claude Code. I'd bet the amount of devs going for a custom harness is vanishingly small, and, well, the API exists for those use cases.
None of that refutes my argument? Devs don't go for a custom harness because said harnesses aren't allowed to compete in equal standing with Claude Code.
Why are you on the side of making the lives of developers worse by giving them fewer choices?
I watched Alien: Earth recently and was admiring the TUIs everywhere. The show takes place in 2120 and they kept up the aesthetics of the original movies. A future full of synthetic humanoids and interstellar ships kept the TUI. I do not remember tablets being in the series, however even those seem to use a similar TUI.
I look forward to my great grandchildren rediscovering the TUI.
I poked around at the code and I like some of the concepts here. Introducing jinjapy as a jinja template with frontmatter allowing me to write python code is useful in some scenarios.
I got most excited about sqlorm (https://github.com/hyperflask/sqlorm). It is way too early to feel comfortable adopting this, but I really like the concepts. I use pydantic a lot, I would love to see the SQL as docstrings as a pydantic extension. I get tired of writing serializers everywhere and would prefer to have a single pydantic model to reference throughout my codebase. Primarily I use Django, but these days am using less and less of the framework batteries outside of the ORM.
As an autistic person with 3 kids with ASD, I get triggered by this bullshit nearly daily. I sincerely hope that whatever "reveal" they have planned does not further harm our ability to access the early interventions that my kids are benefiting from.
Not to be pessimistic, but I’m pretty sure that whatever they do will do more harm than good, given how much of a hot mess the current administration is.
Welp, better hope they don't decide it's a heritable condition that can only be solved via sterilization programs! Because that has absolutely happened before in North America.
I think it's more likely that it will be used as an excuse to actively harm and oppress autistic people and anyone else they can use pseudoscience to try to label as 'damaged' or 'unnatural', rather than merely taking away existing aid.
I was using Windsurf in Pycharm, until I switched to Claude Code / Gemini CLI last week. I am also finding that Copilot with Sonnet 4 is pretty on par to Windsurf.
I like to think of my interactions with an LLM like I'm explaining a request to a junior engineer or non engineering person. You have to be more verbose to someone who has zero context in order for them to execute a task correctly. The LLM only has the context you provided so they fail hard like a junior engineer would at a complicated task with no experience.
I like to think of my interactions with an LLM like I'm explaining a request to a junior engineer or non engineering person. You have to be more verbose to someone who has zero context in order for them to execute a task correctly. The LLM only has the context you provided so they fail hard like a junior engineer would at a complicated task with no experience.
Life is imitating too many dystopian books, movies, etc these days. I think we need to put an end to all creative works before the timeline becomes irrecoverably destroyed.
