Equipment organization
The following are device usage scenarios I’ve compiled based on my actual experience and after consulting with AI.
This is only suitable for my own use and represents areas I plan to explore in the future.
Message notification: gotify + NTFY (Raspberry Pi Zerow)
Automation service: Qinglong Panel (Raspberry Pi Zero2w)
Message gateway: Eclipse-Mosquitto (MQTT) (Raspberry Pi Zero2w; EMQX (Raspberry Pi 3B+))
Network proxy: FRP and VPN (WireGuard) (Raspberry Pi 3B+)
API server or web server: WordPress FastAPI (Raspberry Pi 3B+)
Soft router: OpenWRT or ad-blocking DNS (Pi-Hold) (Raspberry Pi 3B+)
Smart home: Home Assistant (Raspberry Pi 4B (2GB or more))
Cloud phone: LineageOS (Raspberry Pi 5B)
Additionally (I’ll compile and supplement the details later):
Website monitoring, data tracking (Uptime)
NAS Nextcloud
Game Box
TV Box
Notebook Box
All of the above are Raspberry Pi, but you can also use other development boards, such as Orange Pi.
Orange Pi 3B is roughly equivalent to Raspberry Pi 4B, and Orange Pi 5B is roughly equivalent to Raspberry Pi 5.
The difference is that Orange Pi is cheaper and has relatively better performance. Raspberry Pi’s biggest advantage is its rich community resources.
Many open-source projects or use cases are unique to Raspberry Pi.
Graphics cards and AI
As our business has grown,
we’ve found that many things are stuck due to graphics cards and AI limitations. Python used to be all-powerful; now AI is like magic.
Even ordinary people can’t live without AI.
For example, AI drawing, AI speech-to-text, AI digitizers, AI lip-syncing—these are all things I need to use on YouTube, as well as AI-generated short videos.
Here, I want to specifically criticize the AI website Heygen.
They advertise unlimited video streaming, with the text listed first, but the total 120 minutes isn’t clearly stated, misleading users into thinking it’s unlimited.
But actually, the total video length is limited to 2 hours, which translates to about $1 per 10 minutes—a huge rip-off.
It’s much better to buy your own graphics card and deploy truly unlimited AI yourself.
If only I had enjoyed gaming more back then! That way I would have known to buy graphics cards and profited from price increases.
I could have enjoyed gaming more, made money from my purchases, and even if I couldn’t resell them, I could have continued to use AI algorithms as my magical productivity tool to provide content services related to video creation on YouTube and other platforms.
Now, looking back, gaming, live streaming, 4K video transcoding and editing, AI algorithms, 3D animation—everything computer-related—relies on graphics cards.
Relying solely on CPU computing power would be at least 100 times slower.
Furthermore, many projects require graphics cards, specifically NVIDIA cards, not AMD gaming cards.
YouTube channel
My YouTube channel is still not making much progress.
Long videos still have zero views, with the better ones getting close to 100 views, but that’s a rare occurrence.
Short videos remain lukewarm, almost always stopping at 1,000 views, a situation that’s been unchanged for a year.
I haven’t been doing any live streams lately because of network instability, a lack of viewers, and no revenue—it’s a pure waste of time.
Tips
It’s been about a month since my last blog post.
The reason is that I’ve been updating the channel with a lot of videos lately, but there’s almost no traffic.
I’m getting tired of updating videos, so I’m switching to blogging to change things up and give my brain a break. Although the blog is even worse off—almost no traffic—I’ll just treat it as an online notebook.