New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
torrent cloud - would you buy at this price?
I am working on creating a torrent cloud, very much similar to put.io and other such services but cheaper. Would you be interested in buying a plan like this?
- 3TB storage, 5TB bandwidth (will double in the near future), add any torrents to your account (highly seeded will be cached, so it will downloaded instantly) and get 1Gbps download speed. No file uploads for now.
Early price would be 2.5$/month (or maybe 2$), might not be the same after the initial release.
Comments
Depends on a lot of things. Firstly, are you working on it or is Claude?
Can you describe a few of "lot of things"?
I am obviously using the AI agents for speeding things up by letting it review, do security audits and offer architectural improvements, just like any competent dev. I am not "vibe-coding", if that's what you are asking.
what about dmca?
can you compete with pulsedmedia on price?
Ignored (unless it's nasty stuff like CSAM or whatever) but it won't matter much unless you start publicly sharing your download links.
DMCA is used for copyright claims. For illegal (not just copyright infringing) content, you get a court order, not a DMCA request.
Pulsemedia is more of a seedbox with root shell. As stated in the post, it will be similar to put.io, seedr, etc, so you can interact from the web UI or API.
I will add the option to seed (1:1) with same plan.
Correct, I was just stating the obvious that anything like that wouldn't be tolerated.
Can you provide these for $2/m?
I’d totally be in
10 Concurrent Slots
Unlimited Downloads
1TB Max Per Download Size
80Gbps Speed
30 Day Seeding Time
API Access
3rd Party Apps
1TB Permanent Storage
Unlimited RSS Feeds
Unlimited Usenet Downloads + NNTP
Web Player
Dashboard Customization
Stremio Integration
I can't compete with Torbox, if that's what you are asking
.
My question be this, its why I would choose your service if I can either pulsed or torbox. All of them including yours is ai anyway.
From market perspective, can you give me a good reason to chooes youre service ? What separtated you from the others, what feature you had provided that the others dont ?
Alright, fair enough. Maybe some ? Like 14D seeding and Stremio integration as a start
Even linux kernel is using AI for dev, have you stopped using that? Torbox also uses AI. It's not a wise argument to make tbh. Debrid is separate use case, never my intention to compete there.
Stremio and seed is definitely in roadmap.
No, just because now i have multiple storage vps, and can turn them into my own seedbox, using qbit + gluetun with killswitch.
btw from what i see, nowadays most people just pick whatever supports Stremio. If yours does, u def gonna get some users. Plus if u have a lot of cached content..
and for me, the only related service i sometimes use is pikpak. Some dead torrents were cached there. Used it when i needed to find my fav old japanese movie...
You misread. I mean all of them use it including you. So what feature can you have that is unique or stand out?
I'm pretty sure the Linux kernel requires each patch to be extremely well-understood by the submitter, whether or not it's AI. Without seeing the code I can only guess, but I suspect the code screams AI.
So this is basically a debrid service?
Yeah, otherwise Linus is gonna personally spawn in your inbox asking what the hell you just pushed
Even using AI as an assistant is not automatically safe.
The problem is AI can still hallucinate bugs, suggest broken fixes, or generate reports for issues that are not even real.
Xillybus AI assisted review, Eli Billauer, June 2026
Greg Kroah Hartman finding a build failure in that patch series, July 2026
Linus Torvalds on AI generated security reports, May 2026
Greg Kroah Hartman rejecting LLM generated cleanup patches in
drivers/staging/, August 2026So “Even Linux kernel is using AI for dev” does not really prove much.
AI assistant or vibe coding, you still need to understand and verify the code yourself. Otherwise you just end up submitting fake bug reports and broken shit generated by the model.
Where can I get this offer?
You can get all these here(aff). or no aff
Obviously not for $2
No shit sherlock. Using it blindly and using it as a tool are two different things. Everyone is well aware that it can introduce subtle bugs. If you don't read the code, you are being reckless.
But yet again, in a year or two, those issues will vanish since the models are improving at an exponential rate.
Everyone is? What world are you living in?
Heh.
The anti-AI and the doomer crowd will be left behind in the dust. It's a tool, a very effective one if used properly. Being so binary about it's usecase is extremely naive.
I use it for code review, it's great for that.
But you clearly haven't seen the hoards of people who say that their previous applications' bugs are gone because, this time, they told Claude "don't make mistakes".
Just compare the output quality from a year ago. Hell, even a few months ago. Even local models like qwen 3.8 27b is remarkable at it's size.
Sure they're improving, but there's a lot more they need to get good at beyond simply not introducing trivial bugs the way they currently do. Architectural design and avoiding creep are some things the latest models aren't much better at.
"Here's a bug, let me fix it by adding 90 more lines of code instead of considering whether the bug is a result of the initial design being flawed" repeats often enough that you end up with a monster.
You can't rely on them to do everything. Have a proper harness, use skills written by domain experts, etc etc.
Yeah see, that's the problem. You don't even need that unless you're "vibe coding". The best users of LLMs who aren't vibe coders are going to be spending more time with architectural design than LLM configuration.
As amazing as they are for some things, they are still incredibly limited in other areas.
Yes 100% agree.
Dang, are usure about it ? Then why do we have so many SSH clients?
While millennials are watching a donkey feed a hungry fish, at least I’m Gen Z.
I remember the days when I used the first GPT release and asked it to make like 80 lines of Python code. It sucked at everything. Now even a 1B model can perform better than that.