adityapatadia

6 ms·

1.
by adityapatadia 13d ago
We do. MCP works on top of it. I say MCP is just lightweight layer on top of REST and as long as it works for end users I don’t see any reason not to adopt it.
2.
by adityapatadia 13d ago
This. We also thought MCP was useless, but now the workflows by ChatGPT and Claude are so good; it's as good as installing an app from the App Store. We launched MCP for our SaaS and immediately found users using it. API and MCP all ar
3.
by adityapatadia 22d ago
What a downgrade. We were riding leopards back then!
4.
Why current LLM costs are not sustainable(aditya.patadia.org)
116 points by adityapatadia 3mo ago|196 comments
5.
Gumlet CLI and Skills: Upload and manage videos with agents(github.com)
1 points by adityapatadia 6mo ago|0 comments
6.
by adityapatadia 10mo ago
Building https://www.gumlet.com Streaming video is still hard to do for a developer today and we are solving that with scalable and cheap infra for streaming.
7.
by adityapatadia 1y ago
This change finally helped our company move to typescript. Just converted many of services to TS and some are WIP. It’s a big win.
8.
by adityapatadia 1y ago
We use libvips as well sir. We can't overstate how much we benefit from your work!
9.
by adityapatadia 1y ago
We regularly donate to project via open collective. We frankly did not see here due to FB involvement I think.
10.
by adityapatadia 1y ago
Jason, here is a story about how much your work impacts us. We run a decently sized company that processes hundreds of millions of images/videos per day. When we first started about 5 years ago, we spent countless hours debugging issue
11.
by adityapatadia 2y ago
For website hosting, it's okay but not great. We encountered issues when we tried to cache a lot of images. Their CDN storage seems really low compared to Cloudflare and Cloudfront. It results in a really bad hit ratio the moment we tr
12.
by adityapatadia 2y ago
It’s time saas apps realise that they can’t make 2.5x normal license money by just sticking AI to it. In our SaaS we added it for free. We realised that there is no way to sustainably make money off of this in long term. It’s a great featur
13.
How DRM works (video and audio DRM handbook)(howdrm.works)
3 points by adityapatadia 2y ago|0 comments
14.
by adityapatadia 2y ago
Yes, App Platform. Pricing is a little higher but way lower than AWS but it is fully justified. Zero downtime in the last 1 year. With Fly, we had 3-4 downtimes in 2023 in a span of 4 months.
15.
by adityapatadia 2y ago
We left it about a year ago due to reliability issues. We now use digitalocean apps and working like a charm. Zero downtime with DO.
16.
by adityapatadia 2y ago
For image processing, use a proper SaaS dedicated to it: https://www.gumlet.com/image-optimization/
17.
by adityapatadia 2y ago
These servers are indeed job processing servers. They are critical but not milliseconds critical. Cooling, security, monitoring, backup generators, and backup of data all are taken care of.
18.
by adityapatadia 2y ago
3 years is what we planned. Any extra time we get will be a plus.
19.
by adityapatadia 2y ago
we know but it’s actually pretty good.
20.
by adityapatadia 2y ago
aditya [at] Gumlet.com
21.
by adityapatadia 2y ago
Yes, UPS + 2 independent power lines. Our server has a dual power input PSU. If either power fails, the other one keeps it running.
22.
by adityapatadia 2y ago
Nice ideas, but we have chosen a really simple Kubernetes deployment. We only install the host OS (ubuntu server) and then join the self-hosted GPUs as workers in a Kubernetes cluster. No other task is needed and our Grafana monitors if the
23.
by adityapatadia 2y ago
We bought 21. This is just a start.
24.
by adityapatadia 2y ago
Tailscale :) We, fortunately, don't need these exposed to the internet so Tailscale works beautifully.
25.
by adityapatadia 2y ago
Interestingly, we tried that RTX4000 before we decided to buy our own. Yes, ours will break even in 14 months. ($2300 cost + $40 per month datacenter cost)
26.
by adityapatadia 2y ago
I think a single L40 costs $1320 a month on latitude. L40 is also an older GPU.
27.
by adityapatadia 2y ago
$40 per server per month. Includes bandwidth, cooling and internet.
28.
by adityapatadia 2y ago
Probably yes but we just built it to be cheaper. AM5 is on the costlier side and we don't plan to upgrade these machines. Our calculation is we can retire them by the end of 3 years.
29.
We are self-hosting our GPUs(gumlet.com)
67 points by adityapatadia 2y ago|59 comments
30.
by adityapatadia 2y ago
TBH we are at around 15M queries per hour. I am sure our customers don't want us to run on RPi. Btw, it's not only query but billion+ rows which are also there.

More ›