
semisol
👨💻 software developer
📨 nostr.land relay
all opinions are my own.
Commit count, tokens or lines of code are not a measure of useful work.
LoC is sometimes but only if your engineers are competent but even then..
DeepSeek too unintelligent (it has absorbed too much of the Claude pseudointellectualism) and expensive (as I do not use the official provider)
On another note, I genuinely do not get how people use Claude models given that they try to sound smart but aren’t
In order:
- GPT 6.1/5.6 Sol (6.x has tendency to overengineer sometimes, did not try 6.2 yet)
- GPT 6 Luna
- GLM 5.3 Flash
Astra is practically unusable without sub plans and even then limits are too low
GLM 5.3 Flash and GPT 6 Luna. Still too expensive
AI is stupidly expensive
Is this allowed by the policies?
(But who cares about them 😉)
Am I one of the only ones on Nostr who has not given into the vibeslop?
Turnstiles in public transport are dumb
wtf is a napp and napplet and what is even the difference because there appear to be more than 1 definition of it
in other news, LLMs are still stupid as fuck
“What if the CSPRNG doesn’t actually work?” You’re telling me that /dev/urandom is horribly broken and the first thing I worry about is a shitty replay-protection rather than the millions of other things?
🤔
🪨
What you do is, you hide this process from the user, and let the app handle it like normal.
You just say: “Please open your app before <date> to ensure it is up to date.” That is all
replied
It seems that the only place their notes showed up on was aggr, so I did not initially see the notification 🙃
The broadcast relay list is not written anywhere yet but it has all the major relays and like 500+ of them
The real solution is realizing agents don’t work
I agree, best choice
Do not buy DRAMless NVMes for the ones in a case.
While DRAMless drives are fine normally as many host systems support HMB, allowing the drive to map some host RAM to use as cache, the NVMe/USB bridges have no such RAM and your performance will be utter horse shit
Spam
The problem was that they likely closed the tab while the payment was processing. This is due to a limitation of the payment backend.
The issue was resolved.
Eden was merged into Nostr.land.
Due to a limitation of the LN payment processing software used, if you close the tab before it shows the payment as confirmed, it can cause the payment to get stuck.
In this case, just reach out to me. The in-progress rewrite of the backend also addresses this problem.
Note that ranting like this just misleads people and doesn’t help fix the problem. I understand it can be frustrating but we’re all trying to make it work here :)
Could you please send me the invoice reference? I will get it sorted
Much of the costs comes from relays and clients
The total economic cost to process a zap is likely around 100-200 sats
extremely unreliable. also, relays cap filter results at like 500-1000 and many clients don’t paginate so it can appear to “saturate” after some point
Good morning all.
Please dump your complaints about Nostr and relays and media hosting and clients in general
Would you like to fix the load-bearing seam I warned you about, or would you like to learn more about how the new features work?
but, good norning
Please don’t post clanker into kind 1 :(