HN user
I haven’t seen much on WebMCP adoption yet, but I’ve been keen to use it for improving the AX of a web app at work. I’m curious how you all are seeing it be used.
You may not be the creator, but the intent of this repo appears to be at odds with GitHub's acceptable use policies.
For anyone else dealing with recruiters or companies spamming you from your github commit email, it's reportable under information usage restrictions and privacy.
https://docs.github.com/en/site-policy/acceptable-use-polici...
Interesting benchmarks. I'm surprised vLLM is performing so poorly given all the corporate investment/attention it's getting. Is it because the contributions are more distributed than the average LLM team?
Having worked on pricing strategy, this is not what I've seen as the norm in B2B. If you're talking to an early company that is still working on product market fit, maybe, and definitely if you ask for your own features.
If you're talking to larger companies, thing FAANG, then they have a list price and discount levels that can act as incentives, levers or there are other options for inducement. Otherwise, you give the sales team the authority on go-to-market strategy while they are executing individual deals (tactics). Senior sales leaders can authorize some of those discounts and any special inducements or incentives have to be custom written into contracts by legal + deal desk, making them more time intensive and less desirable.
They have one time payments now
Saw Rudderstack doing this via GitHub Sponsors as a bounty program. Wonder if the open collective community might have something to share
From a speed standpoint, I'll say the obvious thing, "it's architecture dependent".
For thousands to tens of thousands of users, most folks need data center with 1+n nodes (w/ larger diminishing returns at 8 nodes). There's a few tips and tricks with that, like pointing CI/CD tools to a smart mirror instead of the primary instance and only mirroring specific build projects. Putting smart mirrors in the same physical location as a primary instance could also help reduce the impact of CPU-intensive requests like git fetch or clone.
Another common issue is apps/plugins. Bitbucket 5.9 added a way to get app diagnostics so infra folks can see if there's a long-running process or other weirdness. Vanilla Bitbucket needs about a 1gb heap (java) and with apps, you want around 4gb (give or take).
If you want a longer read to share, this still has some good info despite being 2~ years old: https://developer.atlassian.com/blog/2016/12/how-we-built-bi...
I'd be interested to hear about the slow behavior you're seeing. There are some tips and tricks for improving that.
Atlassian employee that works with/on Bitbucket here.
We're working on improving that experience. I'd keep a lookout on the next few releases.
Nothing about my comment implies that unless you are trying to be argumentative.
I'd love to see a revised version of this for 2018 (with fewer NSA tinfoil references).
In a past life as a solutions engineer at a SaaS company, I'll address a couple points specific to a unified solution.
Problem 1: Not every SaaS platform has a company with an API to manage user accounts. Even then, I would be skeptical of a company that offered it and didn't offer it via oauth tokens.
Problem 2: Automating the task within the browser also fails when it comes to uniformity. Any company that lacks an API endpoint for user management means you need to interact with a browser or some other hacky nonsense. With that solution comes the problem of understanding the site structure, login forms, and action menus.
Problem 3: Even if you did the above 2, you now have additional points of failure within your offboarding. If a failure occurs in the automation process, is it silent? What if the API changes (not that it should) or the UI?
The best solution is to look for companies that offer the API option or that support SAML.