Recently started playing Go again after years.

Playing 'Go' again after a few years away, mostly against either an AI or my brother (that also only recently started playing). If anyone wants to play a match I would be down (probably only would have time for 9x9 or 13x13 though).

I would be doing it through https://www.funnode.com/

I read a Manga once, Hikaru no Go - Wikipedia, protagonist plays Go,

I found it fun to read, though I was more interested in the character development/interactions/goals/truths, than the game.

Tried playing the game against a computer, a couple times,

But never 'really learned how to play,

Though to myself, looks fun, interesting,

But never spent anymore time on it.

I think the manga/anime Hikaru no Go is how pretty much everyone in the west that knows about the game knows about it. It is a little disappointing as the game is not only the oldest continually played game in history but also the last board game that an AI was able to be built that could play on a professional level (and even then it only won 4-1). Such a fascinating game that is difficult to get used to but also really forces you to think.

Sadly, haven't played it in just over a month as I have been too busy, and even then I had only started relearning the game last November.

I know about it due to geekiness and being aware that the AIs used to beat Chess are less advanced only than the ones used to beat Go.

Go relies heavily on abstract thinking and using it to recognize patterns that might or might not even exist. Lack of abstract thinking is easily punishable which causes it to be harder to 'brute force' your way to victory. This made it significantly harder to develop an AI that could play at world-champ level in comparison to games like Chess.

The AI disagrees with that. The AI has 100% winrate now against the world champions.

What you call abstract just means more lines need to be calculated in a small amount of time. The AI also studies past games of opponents and knows tendencies, enabling it to decide between 3 or 4 different routes and consistently taking the game to the route that the enemy is statistically least familiar with.

I mean, the source even says that "Due to the sheer number of possibilities in Go, traditional “brute-force” search trees which run through all legal variations are not only highly inefficient, but also implausible to implement in a time-constrained scenario."

Seems to me like it is saying that even without as strict of time-constraints that its effectiveness would be difficult if it relied solely on a brute-force method. [EDIT: I also would say that the time-constraints are a part of the game and as such the fact that brute-force methods are inefficient due to said time constraint does indicate the issue as well.] It also doesn't need to rely so heavy on its own capacity for abstract thinking when it relies on looking at numerous games people have played. Compare this with many types of chess AI where they can create professional level AI without having it need to refer to records of human-played games.

I'm not saying that the AI hasn't been able to overcome large hurdles, but I think that it is undeniable that the approach to building the AI shows how different the thinking required is for Go in comparison to games like chess.

Looks like it works with Tor, I would be up for some games. I am not new though, been playing since college.

Machine learning excels at hard to quantify problems