home / future

Optimism and pessimism, both barrels.

I read model releases the way other people read sports scores. Here is my honest read of where this decade goes: what deserves wild hope, what deserves a hard stare, and my personal bet on personal superintelligence.

$ forecast --horizon 2026-2030 · confidence: calibrated-ish
🌟

Be wildly optimistic about

the upside case
likely · 2026-28Computer-use agents go mainstreamAI stops chatting and starts acting: booking, filing, researching inside real software. My February 2026 post called this shift; every release since has confirmed it.
likely · by 2028Open-weight models close most of the gap

Llama-class, DeepSeek-class and community fine-tunes keep eating proprietary advantages. The frontier stays closed longer, but capability parity at the useful layer arrives fast.
plausible · 2027-30AI meaningfully accelerates disease researchProtein folding was the trailer. Drug candidate screening, diagnostic triage and rare-disease matching compound quietly until headlines can't ignore them.
plausible · 2026-29Machines help crack hard mathFormal proof assistants plus LLMs already assist working mathematicians. Expect verified novel results where AI is co-author in substance, not just credit.
already happeningHumanity's Last Exam gets saturatedBenchmarks built to be the final boss keep falling within months. The lesson: measure agents on real economy tasks, because trivia is dead as a test.
⚠️

Stay sharply pessimistic about

the honest risks
certain frictionDisplacement outruns reskillingEntry-level cognitive work compresses faster than institutions retrain people. The gap year becomes the gap decade unless we take transition seriously.
underpriced riskAgent misuse scales with agent usefulnessThe same autonomy that files your reports can run scams at machine speed. Guardrails, audits and permission design become core engineering, not afterthoughts.
structuralCompute and energy concentrate powerTraining runs cost nations, not startups. Without open weights and efficient local models, the future defaults to a handful of boardrooms. Build open.
already visibleSlop drowns signal55K posts taught me this firsthand: generated content is infinite, attention is finite. Verified humans with receipts become more valuable precisely because fakes are free.
genuinely uncertainSelf-improving loops before safety catches upResearch like ai-2027.com sketches recursive self-improvement arriving amid geopolitical racing. I do not claim to know the date. I claim pretending the question is silly is the least serious position available.
signals I watch

Dashboard, not crystal ball.

context window growth

From pages to 1M-token conversations in three years. The next agent frontier is reliable long-horizon memory, not chat length.

open-weight parity lag

How many months after the frontier do open models reach parity at the useful layer? That lag is the democratization metric that matters.

computer-use reliability

Can an agent drive real software five times in a row without a human rescue? Benchmarks are catching up to this definition.

inference cost curve

Dollars per million tokens and watts per task. The half-life of agent economics shortens every quarter — pricing power migrates to whoever runs the cheapest capable private model.

my preparation playbook

What I do while the debate continues.

🚀

Ship daily

A deployed agent teaches more than ten benchmark tables. CareerZen, the project thumbnails and every merged PR here are the practice field.

🧪

Build evals before demos

Without a harness that can say "this broke," autonomy is theater. I start projects with the test bench that will fire me.

📚

Learn the harness, not the hype

Tools, memory, retries, guardrails and permissions — the wrapper around the model determines whether an agent ships. I write about harnesses because prompts don't survive production.

🧾

Keep receipts

Verified numbers, public repos, public writing. In a world of generated claims, a track record you can click through is the most durable advantage.

my personal bet

“Between the utopias and the obituaries sits something quieter and more probable: personal superintelligence. Not one giant mind in a datacenter, but millions of individuals whose judgment, output and reach get multiplied by agents that know their context deeply. The gap between people with a working agency stack and people without one will define this decade more than any single model release.”

ANIRUDDHA ADAK · kolkata

“The future is not something we enter. It is something we ship.”

Ctrl + K open · Esc close · ↑↓ navigate