I was lucky enough to be invited back to the Emergent Ventures Unconference again this year. This is run by Emergent Ventures out of George Mason University, started by Tyler Cowen. This was my second time being invited, I wrote about the first time here. I've received three grants, the latest one being for the JWST API.
The format of the Unconference is something I have not seen anywhere else. It really puts forward this battle of ideas with the freedom to explore without any judgment. This is how it works:
- There are several venues and several time slots available to talk
- A grid board is put up representing these
- Anyone who wants to put a talk forward, writes the talk title on a sticky note
- Anyone who wants to attend a talk, finds one or more talks of interest
- People are not discouraged from attending multiple talks
There were so many different types of people at the event again: authors, researchers, economists, philosophers, scientists, musicians and engineers. With the eclectic group of people who attended the event, the range of topics was incredibly broad.
This format shares some core tenets from the build in public movement, where transparency and openness are key and ideas are discussed on their merit. Discussing ideas openly and being challenged directly on them benefits everyone involved, and with the people in attendance having such a wide background you get an incredible breadth of viewpoints.
As you can imagine, the topics ranged wildly. I put forward two ideas: "AI: work accelerator or career killer" and "Self-replicating AI". The first discussion was around how AI would be a work enabler more than a work killer, but delving into where this might not be the case and what would happen subsequently. For instance, if you want to have your product shot on location in Greece with a blonde model with blue eyes, you can do this now through an AI agency for 1/100th of the cost. What happens to everyone who would have been involved in that chain? The travel agents, the location, the people on set, the clothing manufacturers, the prop providers and people, the models themselves all will have their jobs affected.
The second topic, self-replicating AI, was primarily around an idea I had where you could use software to create software. This evolved into software creating, maintaining and improving software - basically replacing PMs, engineers and testers in the chain. We dove into how you could do this and the implications of it. Since then, AutoGPT and BabyAGI have come around where people are starting to experiment with this exact topic. Of course, the discussion then went to what a GPT or LLM virus might look like. The answer in short is not great at all, and much more like a living virus. Picture the common cold but in code.
Some other talks I attended were:
- "Rants about outer space" where all sorts of space related discussions happened, from someone in her early 20s who has started a space company focusing on better astronaut suits.
- "AI safety" which was moderated by Tyler and Leopold and was very interesting. Basically, discussing how we should be approaching the AI debate with risks and benefits in mind. Tyler has written about his view previously.
- "Funding deep tech" where we discussed how capital can better be allocated for ideas that necessarily take years before any feedback is received or milestone is reached.
With so much discussion my mind ended up racing, so I decided to spend several slots having discussions with people during the discussions. It meant I lost out on some interesting topics, but got to dive a little deeper with individuals.
I've found the Unconference way of running to be such a great approach to knowledge intake: put forward what you want to talk about, go to whatever talk interests you, go to as many or as few as you want. This freedom to roam really opens up your mind to so many different ideas.
After what felt like too short a time, even including an afternoon trip to the National Mall, the conference was over. A few days of great conversations and catch ups with both old and new friends was intensely satisfying. Emergent Ventures are growing every year, so really hoping I get to attend the next Unconference.