5.5 is overcomplicating it.
Where the solution is e.g. changing some oidc auth url, it goes around and verifies and check and builds this and that to eventually change the url, and then write a summary.
It is unable to do K.I.S.S .
Instead of adding just an endpoint, it creates a service, middleware, config reader and finally an endpoint.
LLMs are nowhere near being good developers. The only thing they have is speed. Because of this speed they create the illusion of a good developer, the whoa moment.
Whoa it would've taken me 2 months to implement this.
Yeah but then again you would not make such silly mistakes and you would've reused that oidc client instead of reinventing the wheel every single time.
you will pay 6fold at the very least.
Your 300 messages became 50 or your 1500 became 250 per month.
And you have session limits, daily weekly and monthly limits
Openclaw is unreliable.
I had it running for a few months. It uses up a lot of resources and doesn't provide any benefit other than being able to chat with it via other methods than tui.
because it matters.
Why would you intentionally choose to ignore that fact if it was provided?
I have been using LLMs since August last year, and I know the output they can produce. And I know that the initial output requires refinement in most cases.
And that's coming from someone experienced in Software development.
LLMs in the hands of people who are not experienced lead to skip a proper review process.
Additionally, it's unreasonable to assume one can take a large codebase and will spend hours on examining the code before. It's not only unreasonable but downright ridiculous.
LLMs are a part of reality right now and they're not going away.
Code should be labeled as such.
Not doing that is inconsiderate.
They report content they don't agree with. And if the instance owner shared that, the federated user or even instance gets banned.
Most prominent example is the Ukraine.
Even if you post truthfully and in context, that the Ukraine did support the 3rd Reich with link to the Wikipedia article, you get banned, no warning.
So there is political censorship.
And the fediverse is super political.
The 3 big content areas in the Fediverse are politics, LGBT and nudes/nudes drawings including sadly CSAM.
Blocking (and reporting) an instance because of CSAM I can understand. But that too should happen at an individual level.
It's 1 person, the instance admin, deciding what's good or bad for all their users.
And this is, I'm not sure what, but gives instance admins too much power to moderate content the way they like vs giving the individual user the choice to decide on their own.
Call it whatever you like.
The Fediverse has one problem, concentration of users on few instances, mastodon.social being the largest.
And cancel culture.
Highly politically motivated cancel culture.
What right do they believe to have to dictate to their users what the can and can't read?
That should be solely in the user's hand.
The irony of writing this in HN is ... whatever the right word is
Also, fragmentation and visibility.
It's neigh impossible to find interesting content if you're not on the main big instances.
But React has the largest ecosystem