Continuous-learning AI infrastructure for long-horizon collaborative projects—governed RSI with human attribution. By @MIT scientists. Linkedin: 160K followers
🚀 Funding Round News 🎉 Daice 🎲 Labs
The problem
Today’s AI shines on benchmarks but breaks in changing real world conditions: brittle under distribution shift, costly to retrain, hard to audit, and slow to adapt. We need new hybrid paradigms that integrate today’s models into architectures that continuously learn, generalize across context, and explain their decisions.
Our approach
Daice Labs is computing natural adaptive intelligence—hybrid AI frameworks inspired by principles from living systems. We combine LLMs/DL with symbolic reasoning and bio-inspired architectures to make systems adaptive and interpretable.
News
🚀 We’ve closed an expandable $2M pre-seed round 🎉. $1.5M is secured today, with additional commitments as we hit milestones.
What we're building:
Product Lab
- Advancing/Scaling our composite hybrid AI framework for vertical applications.
- Delivering a hybrid collaboration platform, where teams and agents co-build, learn, execute workflows, and audit provenance.
AI Research Lab
- Accelerating R&D in natural adaptive intelligence, translating cellular and evolutionary principles into hybrid system design.
- Advancing dynamic‑systems methods & cell-inspired architectures.
- Translating breakthroughs into next‑gen systems for non-stationary environments.
Founded by MIT scientists and based in Boston, we are grateful to our investors, advisors, and early partners for believing in this vision.
We are hiring, join us!
An AI can fix a mistake without becoming less likely to repeat it. Real improvement changes how it handles future work.
That requires four clear answers: what changes, what stays protected, how it is tested and who authorizes it.
#ArtificialIntelligence#AIResearch #NeurosymbolicAI #SoftwareEngineering #DaiceLabs
Shared effort built engines that changed humanity.
We’re building Dandela for long-horizon collaboration, connecting people, AI and shared knowledge, with people governing AI’s self-improvement.
Bring your piece.
#Dandela#LongHorizonCollaboration#RecursiveSelfImprovement #CollectiveIntelligence #HumanCenteredAI
AI’s ability to adapt depends on the machinery beneath it.
Memory preserves experience, reasoning tests possibilities, and hardware sets the limits of time and energy.
Building systems that keep learning means designing these parts to work together.
#ArtificialIntelligence #AdaptiveAI #AIInfrastructure #NeurosymbolicAI #DaiceLabs
A correct AI answer can still arrive too late to matter.
Reasoning takes memory, time and energy, so where data lives and how far it moves shape what a machine can do.
Useful intelligence depends on designing logic and hardware together.
#ArtificialIntelligence #NeurosymbolicAI #AIHardware #ComputerScience #DaiceLabs
Every AI answer has a physical cost.
Moving data takes energy, memory has limits, and computation takes time.
Designing intelligence means deciding what to remember, when to compute and when more work is worth its cost.
#AI#ComputerScience#EfficientAI
AI can have the computing power to answer and still be waiting for data.
Memory access shapes how quickly it can respond and how much energy it uses.
Building adaptive AI means organizing knowledge so it arrives while there is still time to act.
#ArtificialIntelligence #AIHardware #ComputerScience #AdaptiveAI #DaiceLabs
More AI agents can mean more ways for mistakes to spread.
Accountable AI needs coordination, checks, and records that show how decisions became actions.
People must retain the authority to set limits, challenge results, and stop the work.
#AgenticAI#ResponsibleAI #MultiAgentSystems #AIGovernance #DaiceLab
Every AI agent changes the world another agent must navigate.
As they share information and act, both discoveries and mistakes can spread.
Building collective intelligence means designing the rules, records and checks that make cooperation dependable.
#AgenticAI #MultiAgentSystems #CollectiveIntelligence #NeuroSymbolicAI #DaiceLabs
What makes an AI team work?
Shared memory, clear permissions and independent checks help turn individual agents into a dependable organization.
Intelligence also depends on how the parts work together.
Read the article: https://t.co/jJAHwpks4I
#AI#AIAgents#NeurosymbolicAI
What makes an AI team more than a crowd?
Coordination divides the work. Governance defines who may act, what evidence is required and when to stop.
Read the article: https://t.co/QxvguHW0H2
#MultiAgentAI
🚀 🎉 🎲
We're thrilled to announce that Daice Labs Inc. is live. A research company pionering hybrid AI systems. Two tracks;
1) our Product Lab is building a collaboration fabric where human teams and AI co-build, execute workflows, co-own and audit outcomes.
2) Our Adapting Machines research lab explores natural intelligence computation for hybrid AI design.
Built by a team of computational scientists, AI engineers, bioengineers, researchers, enterpreneurs, and generalists with roots at MIT, Broad institute, Karolinska institute, BU, and beyond.
Your win powers everyone's upside–shape the odds together.