It is a good time for moonshots.
AI has reached a level where there are transformative projects that could result in huge social good, but require public R&D, consensus & transparency to pull off. Examples: universal tutors, co-scientist/replication systems, remote medical help.
Positive users see recent AI advances with open models making moonshot projects for social good achievable, while negative users worry trust issues and limited capabilities remain barriers.
Current open models are now good enough to pull off some of these projects if scaffolded properly while others (co-scientist) benefit from the AI frontier.
Because of that, this is an area where nations without frontier labs could contribute to the impact of AI for public good.
It is a good time for moonshots.
AI has reached a level where there are transformative projects that could result in huge social good, but require public R&D, consensus & transparency to pull off. Examples: universal tutors, co-scientist/replication systems, remote medical help.
@emollick AI scientists are coming... https://www.nature.com/articles/s41586-026-10265-5
@emollick Yep. Feels like one of the few moments where think bigger is not just conference wallpaper.
@emollick spot on. AI moonshots for tutors & medicine need public R&D & transparency. let's do this!
@emollick We need the universal translator! Technologies that break down the barriers to communication accelerate development of all other technologies.
@emollick every one on that list is already being tried in private. the bottleneck isn't capability or funding. it's that the trust loop resets to zero the first time a model gives a kid a bad chemistry tip.
@emollick Refactoring law systems, optimising govt workflows and procedures
Too many things will be blocked by social inertia rather than the capability itself
@emollick With Opus and other models moonshots are not moonshots anymore.
@emollick Where do you think an average user should start?
The universal tutor has clear institutional buyers (provosts, deans, learning design teams...) who are actively looking for it. At @ARUSTO_AI we work with universities and publishers, but the gap we keep hitting is the same: most can't define what exactly the student needs to be able to do before they build the content, which means the tutor has no north star to personalise toward.
@emollick Moonshots sound great, but the real moat is who controls the bottlenecks — compute, data, energy.
That's where the public R&D should go first.
@emollick AI is finally reaching the point where these ideas feel achievable
@emollick I'm afraid moonshots aren't possible with just Opus
@emollick Open models lower the barrier to experimentation, but public trust, governance, and deployment capacity may become the true bottlenecks.
@emollick Agree on moonshots, but the hard part is credible public metrics + governance. For universal tutors/co-scientists/telehealth, what are the outcome measures (learning gains, replication rates, clinical safety), and what transparency/privacy guardrails would be non‑negotiable?
@emollick 100 pct. This is exactly my thinking as well. As RSi comes on line especially. Further, with open weight models being 90 pct of the top tier and harnesses relatively straightforward for large firms to deploy, I can see no reason to not be all in.