PALMA AI INTENSIVE 2028 · FUTURE LABS
Three intensive days. Six frontier capabilities. A selected room working on real problems with real systems.
THE FORMAT
The programme is built on a simple rule: every block must teach a capability that participants can build, test or operate. If a topic can only be explained in slides, it does not belong here.
Each lab is designed around a frontier capability, a working environment and enough implementation support to keep the room building.
Build a live model of a market, system or physical environment. Generate alternate futures, synthetic actors and shocks, then test a real decision before committing capital in the real world.
OUTPUTA decision simulation that can be rerun with changed assumptions.
Start with an outcome rather than a workflow. Build a permissioned system that decomposes the goal, assembles capabilities and resources, verifies progress and adapts during execution.
OUTPUTAn outcome engine with explicit permissions, verification and stop conditions.
Make a real business machine-readable, machine-verifiable and machine-negotiable. Then release buying agents and see whether they discover, trust, compare and select it.
OUTPUTA machine-facing commercial layer an authorised buying agent can evaluate.
Build from capabilities rather than a fixed catalogue. Let the offer, interface, delivery path, terms and price compose around authorised customer intent and constraints.
OUTPUTA dynamic offer architecture with machine-to-machine negotiation.
Connect digital twins, sensors and shared physical-AI infrastructure. Simulate interventions, define safe authority boundaries, act in the real world and feed the result back into the model.
OUTPUTA simulation-to-action feedback loop with explicit safety boundaries.
Build systems that generate, test and improve solutions, then connect the six labs in one live environment where customers, businesses and physical systems must adapt to changing conditions.
OUTPUTAn integrated Palma Future System tested under changing conditions.
THE FINAL CAPSTONE
World models, synthetic customers, machine-readable businesses, dynamic offers, controlled machine-to-machine transactions and physical systems are connected into one environment. Then conditions change — and the room has to make its systems adapt.
FACULTY MODEL
The selection criterion is implementation ability. No person is presented as confirmed until formally agreed.
Owns the build outcome and must prove the capability before confirmation.
Work in the room, debug systems and keep the practical format from becoming a lecture.
Prepares sandboxes, data, APIs, physical stations, safety controls and fallback environments.
Palma Build Test: every Lab Director must show what participants will have working, the real build path, expected failure modes and how the selected cohort will be supported.
FRONTIER LAB PARTNERS
Technology partners can bring infrastructure, models, sandboxes and engineers. Palma retains control of the learning outcome. No sponsor buys the curriculum.
Defines the problem, capability and participant outcome.
Brings a stack powerful enough to make the capability real.
Builds and operates the system instead of watching a demo.
Founding partners will be announced as agreements are confirmed. No company name or logo is displayed as a confirmed partner without written approval.
ADMISSIONS
Applicants submit a CV or résumé, current role and relevant professional background. Applications are reviewed individually, and applying does not guarantee admission.
The remaining balance is due later by the published deadline. No application payment or programme payment is being accepted through this page today.
Read the admissions process