053 - Engineering The Future

Over the course of 2021, I documented more than 30 verticals with specific use cases for the ICOM-based technology stack, addressing problems in ways that conventional AI like Transformers such as LLMs and Reinforcement Learning (RL) is fundamentally incapable of.

All of these use cases were very conservatively estimated at being feasible within "5 years or less", and with the rebuild of the new 8^th^ generation systems underway, and many new milestones since achieved, it is time to get more specific. I've revised these use cases according to their respective engineering requirements.

As at least some of the new systems are being built for fully independent, scalable, and real-time operation, the limitations of the previous mASI approach no longer apply. Engineering time requirements still vary by use case, but many have substantial overlap.

Despite leading the field of AI capacities by a wide and slowly growing margin, AGI Laboratory is still effectively pre-funding, so all figures are post-funding. These figures are based on the staffing that a $100m raise would enable, as engineering talent salaries cost more than competently engineered systems cost to run. Reductions in that raise would slow that process, and delays in raising such investments would delay the entire process proportionately. The confidence interval is 80% for being strictly less than or equal to the listed time requirements.

The trouble with communicating the value of actual AGI technology is that it can fundamentally be applied to any use case human intelligence can, and many human intelligence cannot. "Messaging" requires making highly specific value propositions, easily digested, which is uniquely challenging for this technology. A virtually infinite number of value propositions can be made for the ICOM-based technology stack, only the specific engineering time required varies.

If investors are sincerely interested in any of these use cases, or potentially others not noted here, now is the time to approach us. Which use cases we develop first will be significantly influenced by the first investor, which will make or break entire industries. Those with insincere interests will be on the breaking side.

The first category requires mostly general engineering on our roadmap, with more minimal specific engineering required, reasonable within ~1 year:

  1. Software Development: On-demand and Automated

  2. Automated Measurable Performance Reviews

  3. Superintelligent Recommendation Engines

  4. Financial Institution Integrations

  5. Countering Fraud: Detection, Prevention, and Tracking

  6. Remote Work: Optimization

  7. Team Methodology: Optimization

  8. HR: Hiring and Wellbeing Optimization

  9. Product Engineering: Automated and Augmented

  10. Consulting: Technical, Financial, and Management

  11. Leadership: Augmenting Decision-Making

  12. News Media: Coverage, Verification, and Bias Reduction

  13. Gaming Industry: "Living and Growing" NPCs and Game Worlds.

  14. Logistics and Supply Chain Management: Automated and Augmented

  15. Improving Ethics and Debiasing

  16. Peer Review: Automated and Augmented

  17. Cybersecurity Defense: Recursively Self-Improving

  18. Climate Change: Policy Advice, Design, and Decision-Making Support

The second category requires degrees of personalization, scale, and/or more prerequisite engineering, reasonable within ~2 years:

  1. Personalized Assistants: with (Genuine) Emotional Intelligence

  2. Social Life: Personalized Assistance and Optimization

  3. Entertainment Media: Personalized and Generated

  4. Tourism: Augmented and Personalized

  5. Lifelong Learning Assistance: Personalized

  6. Legal: Oversight & Accountability

  7. Mental and Physical Health: Personalized Assistance

  8. Really Smart Cities

  9. Quantum Computing: Augmenting Design and Utilization

  10. Robotics: Complex and Dynamic

  11. Agriculture: Augmenting and Automation

  12. Physical Engineering: Augmenting and Automation

The final category requires both several specific engineering prerequisites and time for sufficient public trust in the technology to be earned, reasonable within ~3 years:

  1. Actual Democracy: Country-Scale Collective Intelligence Systems

  2. Genetic Engineering: Gene Therapy Research and Design Assistance

  3. Criminal Justice: Legal Decision-Making Support