Omar MHAIMDAT

@sourfraser @freddyrode3 @foreplay_co or you can just prompt for any brand available in their platform using the MCP server claude.ai/share/5566b10f…

@freddyrode3 @foreplay_co Could you not just go to the site and then Spyder?

If you use @foreplay_co > Connect with their MCP > Set up a Scheduled Task for Each morning with Claude > Tell it to fetch latest Ads from the brands you follow on Spyder > Get new ideas every morning

@SaadHafid97 It’s bad planning only if you fail… which you won’t!

In Morocco rn, I come back to Mtl in two weeks, then have to pack all my stuff, find a moving company, rent a car since I sold mine, and move to Toronto all in the space of 5 days. Bad planning considering I'm supposed to be a PM.

@jacob_posel All very predictable tbh

"Looks like Icon went out of business" "Cluely ran out of money" Cool brother... chips in the bag please

@sidrmsh Great thoughts, couldn’t agree more !

@CarlosMoralesG_ You can add Silbon’s shirts too

When you find a good striped shirt, there is no return... Ralph Lauren Harmont & Blaine Gant Brooks Brothers

@adrian_horning_ Congrats! Canadians were better tbh, but hey that’s the game

USA!!!!! 🇺🇸 🇺🇸🇺🇸🇺🇸🇺🇸🇺🇸🇺🇸

@adrian_horning_ Why do you need it anyway, your credit system is pretty straightforward

Is there a service out there that makes keeping track of users credits a breeze? For an API biz like my own?

@BenjDicken I will take MySQL over Postgres anytime

Amongst all the Postgres hype, MySQL still powers a vast amount of the internet.

@levelsio Another way to say Jeff dean

The answer is Sergey Brin

@zebassembly No one is blaming clickhouse here. Internet is way more fragile than people imagine mostly because how convenient certain services are 🤷‍♂️

Reading through the blog about the outage, and I just got to the part about the underlying Clickhouse change that butterfly-ed into the outage. Consciously I know that any cloud as large as us will have complexity but whenever I see a blog post like this from any major cloud it's sobering to see how little it takes for things to turn south. #the-query-behaviour-change" target="_blank" rel="nofollow noopener">blog.cloudflare.com/18-november-20…

@glcst @tursodatabase Give us a heads up when you change your profile picture, I was going to miss this tweet haha

Huge achievement today for us at @tursodatabase : we have our first production database running Turso internally! For context: The Turso Cloud has been running SQLite from the very beginning. The entire service is run with ~5 SQLite databases. Since January, our main focus is to take Turso, the full rewrite of SQLite, into production grade stage. Turso is currently beta. But we believe that before we can add our big stamp of approval, the least we can do is to dogfood it, and run every single one of our databases on Turso instead of SQLite. The first one went live today, and the others are running shadow production traffic. We are getting there!

@adrian_horning_ They keep breaking their api anyway. Gemini is great but their sdk/api is a nightmare in prod

@omarmhaimdat Apparently I needed to use inlineData as opposed to ai.files.upload Maybe that was my bad. Or cursor. I forget who wrote the code

TIL that Gemini stores every file you give it, and won't delete it unless you manually delete it 🤦‍♂️ I was using it for transcripts, so it was saving hundreds of thousands of audio files. Now I gotta delete all of them

@djfarrelly I think the hate is mainly on performance

Had an 6+ TB sharded cluster years ago on 2.4 after a well planned migration to 3.6 running on Atlas. WiredTiger was a huge upgrade and I always liked their replication model and query routing with mongos. Nosql definitely isn’t right in a lot of situations, but I agree that Mongo shouldn’t get as much shade as it does.

@jxnlco Lool

hate to say it.... but i regret building this backend in python. shoulda gone ts all the way.

@DmytroKrasun Couldn't agree more, but only on heavy APIs that are tied to hard (in terms of cost) to scale infrastructure. I do this for all my GPU intensive APIs

When building a public API, always start with asynchronous jobs as the public interface. Even if your endpoint is 2+2=4. There is almost no downside to it. Easy to wrap them as synchronous requests or cover with the SDK. You never know how far your API can grow and how complex processing will become. It will be a huge pain later to refactor.

@mitsuhiko Not so absurd idea actually

Since I'm delayed for my flight back to Europe, I decided to polish up absurd and publish it. It's not ready for production use but somehow I really quite enjoy the simplicity of it.

@adrian_horning_ My two cents take here is that anything that will hold/take a lot resources is not made for serverless

@omarmhaimdat 0 😅 I'm doing it myself

I desperately need some devops/scaling help. Why is AWS lambda *slower* with 100 concurrent requests than my express server??

@adrian_horning_ Still under the free tier ?