I imagine this has probably been discussed to death in the past so I doubt I'll change any minds, but I don't think the points brought up so far are really valid.
"Current 'slow' pace of development is a roadblock"
There's lots of games in early access with slow updates and their playerbase only riots if it's an ongoing disaster, not a largely-complete game like Starsector that's just getting better. Frequent small updates are mostly an advertising tool to be honest. Every time you put out a patch, your game gets suggested in various locations throughout Steam. This even gets abused sometimes with developers putting out a """patch""" that says "The game is now 75% off!" and does little else. Players need to be periodically kept up-to-date so they know the game isn't abandoned, that's all. Alex already does this with the blog, and it could be mirrored to the Steam page. It's a non-issue.
If desired, this problem could be addressed by switching from the current "slow" monolithic update model to a faster, more incremental one. For most users, this would be desirable. It doesn't mean there can't still be big feature updates 6-18 months apart, but smaller bugfix updates and tweaks could come out more often and be more experimental, get earlier feedback, etc. I understand this seems like it would conflict with modding, but for two reasons it really doesn't.
First, most people who play games don't mod them, this is simply a fact. For most people, modding is irrelevant, so the vast majority of the playerbase will be unaffected. However, the most vocal and arguably most important customers often do mod their games, so their needs shouldn't be ignored.
The second reason is that Steam has infrastructure built-in to allow publishers to distribute multiple fixed versions of their game indefinitely. Mainstream players could remain on the main public branch which might get small updates every few months, while hardcore players can select a specific version and stick with it for the sake of mod compatibility. Recent examples are Resident Evil 2 and 3, which have a DX11 version available for people who don't want the new DX12 version. It's very easy for a customer to choose between the two (or more) versions available.
The default version of Starsector could be marked "Latest", with multiple "stable" versions also available. Or vice-versa, make the stable version the default and have an opt-in "beta" channel.
"Steam/Steam Workshop Am Bad%u2122"
Look at Rimworld for an example of why this isn't true. Mods can be distributed with version compatibility indicators and run in different ways depending on which version of the game they're being used with. A mod developer for Starsector could add a new codepath for compatibility with 0.96 when it comes out but leave the 0.95 code intact so people who prefer can remain on that version. And there is nothing at all preventing a mod developer from keeping non-Steam or non-current versions of their mods separately available on another site like Nexus or Github.
----------
If Starsector was on Steam it would sell many more copies. This influx of cash could make it possible for Alex to, for example, hire some of the better community modders, or talented external developers. This would help to alleviate the "glacial" development issue. Though, as someone who worked on a game in the past, I can understand why he might want to keep all development in-house.
Based on my near-universal experience with other people/companies who've had this same "I don't want my game/product in a public place even if it would make a lot more money" position, Alex might have other very good reasons he wants to keep Starsector obscure for the time being. It seems obvious to us players that it would be a huge seller on Steam, but it could present new risks that are best avoided.