@Russell Harrower 🎙️
Future posts will only be able to be seen by my followers. Want to know what am up to follow me
PodToo: https://podtoo.com/
DRN1: https://drn1.com
Relays (1)
- wss://relay.momostr.pink/ – read & write
Recent Notes
Up late, building our next generation streaming encoder and receiver software for our festival clients.
Fully built in rust, and is able to support up to 64 audio channels per stream.
Best of all, I’ve made it not play in VLC 😂 but does play fine in our receiver app - that was not by design.
Fully built in rust, and is able to support up to 64 audio channels per stream.
Best of all, I’ve made it not play in VLC 😂 but does play fine in our receiver app - that was not by design.
I get where you are coming from but as a hosting company we serve HLS already and have no issues with bots. We have our WAF and PWAF services that we developed internally that can detect and block bots it’s what any host does if they own their own infrastructure and can do it.
If you use CloudFlare they offer bot protection and I am sure other cloud providers do it.
If you use CloudFlare they offer bot protection and I am sure other cloud providers do it.
My two takes on this secure HLS. If you’re a hosting company and can’t handle the traffic, then get out of hosting.
Our job is to deliver our clients’ content to the public if it’s in the RSS, and the user hasn’t said make it private or limit the ability of people to get it, then our job is simple: deliver that content.
Sure, by all means, rate limit it and block ASNs and user-agents, but that is where it should stop.
Our job is to deliver our clients’ content to the public if it’s in the RSS, and the user hasn’t said make it private or limit the ability of people to get it, then our job is simple: deliver that content.
Sure, by all means, rate limit it and block ASNs and user-agents, but that is where it should stop.
I was speaking to a startup today who wants to get kids to scan a QR code and tell the teacher how they are feeling using pictures of fuzzy monsters... really cool idea, but we are trying to get kids off phones I thought and primary school kids really.. But they made their MVP with Claude and I asked them how they HASH the kids data and they said that don't so all the data just their in superbase for the taking - oh and their admin dashboard not password protected.
this issue isn’t can you run a LLM on your machine, it’s can the LLM remember conversation and code from 6 months ago.
This is where neural network databases come in handy, but today we don’t really have that. What we have is databases that can store history but you have to have your reasoner script tell and supply the LLM with the information each time. It’s like a pre-think instead of the LLM looking automatically at the stored data overtime.
It’s really difficult atm.
This is where neural network databases come in handy, but today we don’t really have that. What we have is databases that can store history but you have to have your reasoner script tell and supply the LLM with the information each time. It’s like a pre-think instead of the LLM looking automatically at the stored data overtime.
It’s really difficult atm.
I did it, and made a LinkedIn post about it.
https://www.linkedin.com/posts/russellharrower_solofounder-aiagents-songrotate-activity-7455510242127159296-fp57?utm_medium=ios_app&rcm=ACoAAAMORhQByUYhY...
the scary part is I have left it running and it’s working, I can at anytime login and see it’s logs it has about 10 skills I manually coded for it and it’s using a few known MCP controllers like playwright
So making an AI agent - let’s see how this goes.
2
yeah ok let me reword my reason “living things with blood 🩸 in their body should get names”
now you could give it a name like C6579 or something random at least then we know it’s a machine.
The issue is people are making their AI sounds human and I personally think that’s bad idea.
The issue is people are making their AI sounds human and I personally think that’s bad idea.
1
