tschellenbach

AI has some of the craziest up and down cycles of tech I've ever seen

Definitely looking at using them, you do wonder if such narrow focus is defensible as a business? Probably not.

MongoDB does seem to be in a rough spot. Their pitch is an easier to get started with database, which longer term gives you problems and higher costs. And since we all have AI agents, why would you select that for any use case?

This post is just AI slop. The details of these contracts aren't disclosed as far as I know.

You need to understand the contracts, the acceleration of demand, how the various inputs into supply scale (energy, chips, data centers) etc to say something sensible about this.

New JSON is amazing, and SIMD will be big for json, audio/video etc.

Every release CPU load becomes a bit lower. Love it :)

It's fascinating how well the system recovers. If it works, companies that do remote work will outperform those that don't. And the reverse. If many companies all opt-in to remote work and output declines inflation will correct that.

"Startup, parenting, hobbies, friends. I think you can do 2 of these well. But you'll have to quite light on the others."

Someone on X, don't recall who, but it resonated.

Some good data on this topic: https://ourworldindata.org/co2-emissions

Dutch living in Colorado. You can spin stats any way you like. But if you look at the last few waves of companies. Machines for making chips, manufacturing chips, designing chips, search, social, mobile, AI.

If this was a soccer match the score would be 10-1 (yee ASML). Not only that the winners in these industries tend to compound and grow.

There are many great things about the EU, that are outside of this competitive ecosystem. But I don't think sticking our heads in the sand is really a great idea.

Logical replication solves this. You roll the cluster, downtime is minimal. like 60s maybe.

PgDog, Neki, multigres, awesome to see. And yes this is the main issue with postgres. Well this and not having index hints, looking forward to 19

Imagine you have a SAAS app. Microsoft and Apple are customers. You have a table of devices. Type is either windows or osx. How does postgres know how to handle the query devices, where type = OSX properly? How does it know that this matches ~0 or ~100 of rows depending on the customer?

This is the main thing the planner doesn't handle well. Postgres was built before SAAS was as big. You have different distributions per customer, and thousands of customers. In most cases the query planner will guess right, but sometimes it will fail and scan millions of rows.

We have a stack similar to stainless internally at Stream. Might open source it if there's demand.

I think generally we are going from vector based search, to agentic tool use, and hierarchy based systems like skills.

postgres is great, but its not great at sharding tables.

Google, Spacex, several startups are all doing this. The best people in their fields think it might be viable. I'm skeptical as well, but you do wonder if maybe they are right and how exciting that would be.

Back in the days, I did the first 2m in ARR at Stream myself, it was kinda hard :)

Nice to have a team in place these days, but I still show up for the largest deals to support the team as needed. (140 person company, i think this always stays part of the founder tasks)

Feels like they were first in the space but then somehow Ramp ran away from dev with a higher dev pace. Fascinating to see.

It's cool how lightweight it is. Recently added support to Vision Agents for Pocket. https://github.com/GetStream/Vision-Agents/tree/main/plugins...

Vector & BM25 on Turbopuffer. (see https://github.com/GetStream/Vision-Agents/blob/main/plugins...)

They could build something like Lovable but with better design/frontend defaults.

10 years ago we started out with Python. We switched to Go probably 8 years back. I think my little startup would have utterly failed without Go. Thx google :)

But yes Enums are so much nicer in Kotlin vs Go. That's true, it doesn't impact productivity much, but he has a point.

Well.. there are many fast growing companies that provide UI + APIs for certain components of your app. Sure you can build things easier in-house, but the opportunity cost of doing so also went up. Supabase, Stream, Clerk, Stainless all growing very well.

Why hate on someone doing their best and building.

I wrote this forever ago in AI terms :) https://getstream.io/blog/cursor-ai-large-projects/

But the summary here is that with the right guidance, AI currently crushes it on large codebases.

It depends. For AI to work for large projects (did a post on this forever ago in AI terms. https://getstream.io/blog/cursor-ai-large-projects/)

But you need: a staff level engineer to guide it, great standardization and testing best practices. And yes in that situation you can go 10-50x faster. Many teams/products are not in that environment though.

Tend to agree here. The slowdown here has more to do with the financial ecosystem. IE less capital available for some companies, higher salaries and a changed approach to work.

The AI wave didn't start yet. Will hit in 26/27

The title here mostly doesn't match the article right? Quote: "But unlike the capital growth tax, capital gains tax will, in principle, only be levied at the time of realisation. This is usually when the relevant asset is sold, but also when immovable property exits Box 3 for another reason, such as emigration."

Feels like this should be solved at the iOS/Android/Windows/Sony level. Parents should configure settings in 1 place. Not once for each game.