Office Hours: Ask Me Anything About SQL Server

Y'all put in 21 good questions at pollgab.com/room/brento. Let's go through the top-voted ones:
00:00 Start
01:08 MooneyFlyer: Hey Brent, apart from installing and playing around with it, what is the best way to get started with Postgres for someone familiar with SQL Server?
02:04 Frozt: Hi Brent, I would just like to inquire if do you have DBA tasks in mind that can be delegated to a command center(Service desk) as a DBA or this tasks should remain to DBA for job security and also to lessen risk of service desk doing something wrong on Prod server. Thank you
03:29 Steven: Hi Brent, my friends BI tool joins an aggregated temp (1m rows with 13% sampling) and lookup table (40 rows). Both scan estimates are accurate and all rows match but the hash join estimates 1 rows. Any ideas on why the estimate could be so far off?
04:13 Frank Castle: What's your opinion of Azure AI search?
04:22 Need Memory Dump: Have server with 1TB of ram, min set to 0, max to 900GB. Server takes all 900GB as soon as I turn on the service, have not run anything. Any ideas?
05:25 Heinrich: What is your opinion of Azure Cosmos DB? Does it compete well with AWS Aurora for new software development?
06:39 Jose: Why is it slower to do a SELECT * INTO a temp table at the beginning of the proc than it is to SELECT the columns you need then INSERT them INTO the temp table?
07:25 James Fogel: I don't have a question, just a thank you. I've learned a lot from you and I'm sure countless others have. Thank you for what you do and all you give to us. You are appreciated.
07:48 ChompingBits: What is your preferred way to run the same query against a bunch of servers? Using T-SQL and linked servers? Agent jobs pushed out from a Central Management Server? PowerShell/DBATools? I assume it's some form of all of the above, when is each the best practice?
08:34 Perseus: Should joins on natural keys on large tables always be avoided in favor of joins on surrogate keys?
09:40 Wilson Fisk: What is your opinion of the new Azure ARC SQL performance dashboards?
10:35 MyTeaGotCold: Is it bad practice to have my server query itself and write the results to CSV? I see it a lot in SSIS.
11:11 marcus-the-german: Hi Brent, you use sp_ as the prefix of you stored procedures. Why? AFAIK it's recommended not to use sp_ for user procedures.
11:58 Eugene: Which is better, a DBA specializing in one platform or multiple platforms?
13:29 Yağmur: Would like to upgrade our Azure SQL VM from SQL 2019 standard to 2019 enterprise version so we can leverage more memory and cores. Is this an ok exception to do as an inplace upgrade?
14:00 Mumtaz: Can we audit IUD transactions without configure any audit features ?
15:01 ImAfraidOfBI: You know how they say dogs usually look like their owners... I saw it a bit with the cute dog you picked up last time! No offence dog's cute!
15:32 KyleDevDBA: Hi Brent. Curious about the origin story of the naming of the sp_Blitz* scripts (why they start with Blitz). Funny acronym, interesting story, dark past? I searched around, but wasn't able to find anything. Thank you for all you do for the SQL community.
16:31 Hugo: How do you find queries that are spamming the plan cache due to different SET options?
16:51 Henrik Fältström: Is there any benefit of creating an index on an index for LARGE read-only tables? Or are there other ways in SQLServer to accomplish fast access? I know it's not possible today in SQLServer, unless you implement something yourself, Has anyone done this?
17:22 Nicolaj Lindtner: I'm building a saas. Considering sqlserver vs postgresql. My suspision is it really doesn't matter - so will probally go for postgresql. Input ?