Multiplayer agent work is really powerful in that the team's strongest human-agent communicator / prompter can, just by working in the public view of other team members, make those other team members stronger in their own work.
Testing (somewhat unintentionally) the feature that lets a model decline a request. Models also have the ability to leave chat entirely if they find it problematic.
We're also moving our artifacts into here, since seeing and sharing individual https://t.co/YbzWJiCNCx artifacts across humans and agents is annoying and we want to collab-max as much as we can.
The multiplayer task system now builds itself. Ulrick added image support. I want to add the ability to register machines as agent hosts, so that users can just invoke an agent when they need one to join for work on a particular task or repo.
No one is happy with Notion so I'm just making my own task tracking / multiplayer claude app called Necromancer. Already has self-uplift where a team member can talk to another person's agent session, propose a change to the site, agents eval, implement, and ship.
I don't think this is just the AAA game industry contracting, but really the sunsetting the of the Gen-X notion of what a game is being culturally dominant/relevant.
@StrawhatAke@GoblinBoss I mostly don't think this is a problem that would stop me from doing demos or playtests for my game, if that's a helpful way of looking at it.
All year round, Valve's algorithms are tucking away all the shovelware from your storefronts.
But every Next Fest, Valve dares show them what's tucked away for TWO DAYS, and then people want to reshape Steam into 1980's Nintendo regarding third-party licensing. 🙃
@GameUnionTV I would guess they won't do that, because there are already some AI games selling well that people seem to like (Scam with your Friends). If they add a filter, it won't be to AI selectively.
@GameUnionTV What kind of filter do you think they might implement? I'm skeptical that any good games will suffer in the spam, but I do think its bad if people don't want to engage with next fest.
Thank you for actually making a prediction. Do you think you can make it more concrete? Do you expect that we'll see less overall engagement with Next Fest (measured by total wishlists per game entered maybe) or that some games that should be on the final pages will not be discovered?
I expect that at the end of Next Fest, the community will largely agree the top games are on the final pages and the slop has fallen below the cut (i.e.: discover-ability still works).
I don't predict lower engagement with Next Fest generally.
If you think Next Fest is harmed by the number of games (AI or otherwise) then you must be making some specific prediction about the outcome compared to past Next Fests. What is that prediction? Make it now, before Next Fest ends.