--- title: "2026 Predictions" description: "I'm late on this one so I've seen everyone else's. I'll try my best to make some bold predictions I've not seen elsewhere" date: 2026-01-16 published: false tags: - predictions template: blog-post --- I'm late on this one so I've seen everyone else's. I'll try my best to make some bold predictions I've not seen elsewhere ## more open source, less open contribution I predict that this is the point in time that we see an explosion of open source, but more projects than ever going the way of sqlite, Livestream, android, unreal engine, aesprite. It is getting so easy for agents to generate 10k lines of code in a few hours. This makes it easy for folks who have ideas they want to create to create them. It also makes it much harder to accept outside contributions when they make such massive changes. Thirdly it will be easier for folks to just make their own fork to do what they want. Right now it's hard to name 5 big open source, closed contribution projects by the end of the year there will be 5 more that are recognizable. I think more open contribution will go the way of [[ whenwords ]], spec driven development. This is very high level. Very easy to understand a large portion of the codebase very quickly. Very easy for anyone to understand. While it's all written in plain English and not code it has a way of being deterministic, and does not skip on software engineering principles. ## 2026 is not the year to build hardware The time will come hardware will be cheap. I hold a lot of hope for this, but 2026 is not that year. This really isn't my area of expertise what do I know other than I have gaming PCs, laptops steam decks, and servers in my house. I predict none of them get an upgrade with new parts this year, but there will be a year in the future that upgrades are so cheap that it's hard to pass on. Sites like serverpartdeals will be flooded with good drives from failed data center plans (so many in place right now not all of these companies will survive, not predicting any mass bubble pop here, that's above my pay grade). But 2026 is not that year. ## Agents will overwork us There is going to be burnout this year, people are going to be running point on 12 agents for most of their day and be plain ass burnt out by the end of the year. This shit feels different and its hard to see right now. Most of us can feel programming work, the stress it brings, the joy it brings. We have a decent sense of being almost there, just 10 more minutes to wrap up. It's taken us a long time to hone this craft. We don't have a sense for an agent almost being there. The work we put into planning for agents feels different. It doesn't _feel_ as hard, but it engages a higher level of thinking that allows us to see the forest for the trees. It can let us work on more things at the same time. I predict by the end of the year there will be podcast episodes talking about burnout, and specifically feeling burnt out on planning for agents work. ## 2026 is to ai as 2012 was for js frameworks This year we will see tooling and technique churn like crazy. I think there are already some clear winners with copilot, cursor, claude code, and opencode. I think these will remain in their respective camps as winners and do well. What I think is going to churn a lot is what we hook into them, whether its mcp, lsp, or clis. Currently as I write this we have an issue managing context, getting the right context to the agents, managing bloat and token count. There's a lot of talk about memory. I think the first half of the year we are going to see a lot of tools that hook into the agentic tools to give them more memory, the right context in the session. There's also an issue around getting the agents to run longer. Right now we're seeing things like the Ralph loop where agents keep spinning on the same task. Tools that help agents run longer without getting stuck will be hot. The type of things that are going to have new things come out each week will be tools that... * Manage context * Get agents to run longer * Add skills * Hook in via CLI or mcp Right now I think we have some clear winners around the parent tools, but a layer underneath is ready for a lot of churn. Buckle up. ## Results I'm tracking 2026 Prediction Results