Also some are way more achievable by software-type engineer-types and the financial associates in their ecosystem, due to extreme familiarity with that particular landscape. Some also require a little more commitment than starting a small software company.
If I was going to split the combined vision into only two categories it would be like this:
- Applying machine learning to robotics
- New defense technology
- Bring manufacturing back to America
- New space companies
- Climate tech
- A way to end cancer
- Foundation models for biological systems
and then the less moonshotty efforts:
- Using machine learning to simulate the physical world
- Commercial open source companies
- Spatial computing
- New enterprise resource planning software (ERPs)
- Developer tools inspired by existing internal tools
- Explainable AI
- LLMs for manual back office processes in legacy enterprises
- AI to build enterprise software
- Stablecoin finance
- The managed service organization model for healthcare
- Eliminating middlemen in healthcare
- Better enterprise glue
- Small fine-tuned models as an alternative to giant generic ones
Interestingly, #1 rose to the top of my list well over 40 years ago when I had a chance to do a little machine learning to guide automated systems. Was very lucky to have such powerful advanced equipment under my complete control in the laboratory at such an early time. Needed custom gear to bump it to the next level though. Figured all kinds of people would be doing things like that once "personal" computers were no longer a rare curiosity.
The remaining things in the first group are some other things I (and I'm sure many others) have had in mind since before personal computers became accessible.
"Too bad" my ambition has grown with age and it would take about a $10 million company to build my prototype hardware, and that's before any deployable machine learning can commence.
So it's been an interesting 43 years keeping in mind how I would apply automation and machine learning to almost everything all the time, and refining my intended approach for a greater number of decades the earlier I had the idea.
I'd move "Foundation models for biological systems" to Moonshotty efforts. Foundational models for biological systems are hard, and honestly I think require a degree of patience that is at odds with VC funding.
Also some are way more achievable by software-type engineer-types and the financial associates in their ecosystem, due to extreme familiarity with that particular landscape. Some also require a little more commitment than starting a small software company.
If I was going to split the combined vision into only two categories it would be like this:
and then the less moonshotty efforts: Interestingly, #1 rose to the top of my list well over 40 years ago when I had a chance to do a little machine learning to guide automated systems. Was very lucky to have such powerful advanced equipment under my complete control in the laboratory at such an early time. Needed custom gear to bump it to the next level though. Figured all kinds of people would be doing things like that once "personal" computers were no longer a rare curiosity.The remaining things in the first group are some other things I (and I'm sure many others) have had in mind since before personal computers became accessible.
"Too bad" my ambition has grown with age and it would take about a $10 million company to build my prototype hardware, and that's before any deployable machine learning can commence.
So it's been an interesting 43 years keeping in mind how I would apply automation and machine learning to almost everything all the time, and refining my intended approach for a greater number of decades the earlier I had the idea.