Hacker News Reader

Comment by esafak | Hacker News Reader

Does anyone have recent experience with both? I'm in the market, for a startup, and I need to decide. My priorities are the data warehouse (BigQuery vs. Redshift) and ML (Vertex AI vs. SageMaker).

This article is three years old. Vertex did not even exist then.

View on HN

Don't know enough about Vertex AI, SageMaker is great though. But on BigQuery v Redshift: Having used both over many years, BigQuery is the clear winner. It's a much better, much more modern product.

BigQuery is also a central part of GCP, with almost all services having integrations to it (PubSub, CloudSQL, Bigtable, etc). It also seems to get a lot of resources and is continuously improved. That is not the case for Redshift, which is not as important to AWS.

I've worked in AWS before and also in a company that uses GCP.

Funnily enough, I would recommend AWS over GCP in every single category except for the two that you mentioned.

Although Snowflake is my choice for DW and it is cloud agnostic

Choose whoever gives you the biggest startup credits. The UX and UI experience is secondary to that imo, and the other technical differences aren't something you will care about as a smaller startup.

I would choose Redshift over BigQuery but that's because I know how to use it. Actual skill at databases is increasingly rare, so BigQuery wins this one for most. Watch your bill.

One of the 2 never deprecate anything and ensure your workload will run untouched for 10y or more. The other one will force you every 6 months to update sdk, api endpoints or even kill a service you need in weeks. I let you guess who's who

Reply on news.ycombinator.com