eladgil

Hmm, which verticals did I do this in?

The majority of companies listed I did not invest in, as well as in some cases areas I have not invested in at all

I am actively involved in AI and have been for a few years, so this both gives me insights and obviously conflicts. My hope is to write things that are useful vs just shilling as I would lose credibility otherwise

I agree this is a potential outcome. One big question is generalizability versus niche models. For example, is the best legal model a frontier model + a giant context window + RAG? Or is it a niche model trained or fine tuned for law?

Right now at least people seem to decouple some measures of how smart the model is from knowledge base, and at least for now the really big models seem smartest. So part of the question is well is how insightful / synthesis centric the model needs to be versus effectively doing regressions....

I think this is already true for things based on diffusion models - e.g. Midjourney, Pika etc

For LLMs, Character and ChatGPT are arguably two vertically integrated consumer apps (with some B2B applications for ChatGPT as well)

Would love to get any views / counterpoints on these topics.

Eg which AI infra approach thrives relative to the big clouds?

What new architectures are needed for agents vs transformers?

Etc...

Lots of stuff to figure out....

I forgot to list Bitcoin as another one that launched with no VC money :)

It is remarkable how many of the tech companies that have lasted the longest and got the biggest started off so lean... (although BTC of course is not a company)

Braintrust also provides eval, so nice tie in across both...

Agreed on ML being in production in a lot of places and has been quite valuable, particularly for large incumbents. I write about this a bit here: https://blog.eladgil.com/p/ai-startup-vs-incumbent-value

I think the difference this time is the types of capabilities provided by transformers vs prior waves of AI are sufficiently different to allow many more types of startups to emerge, as well as big changes in some types of enterprise software by incumbents - in ways that were not enabled by pre-existing ML approaches.

100% agree the theory on AI is old and actually dates back to the early days of "cybernetics". But the real difference is at what point do we considered it sufficiently reduced to practice? I chose GPT-3 but undoubtedly people can point to earlier examples as glimpses of what was coming for sure.

The bar for what is "AI" keeps moving. For example plane autopilots would be "AI" in the 1980s, the ability for a machine to win at chess, go, and other games etc.

Definitely not my intention to forgot or denigrate the past. Obviously all this exists due to deep learning and prior architectures. What I have been running into is many people and companies are interpreting this as "just more of the same" for prior ML waves, when really this is an entirely new capability set.

To the (bad) analogy on cars versus planes - both have wheels and can drive on the ground, but planes open up an entirely new dimension / capability set that can transform transportation, logistics, defense and other areas that cars were important, but different enough in.

I agree that most startups fail. Only a small handful of companies at a given moment are the "breakouts" that are clearly working, and those are the most derisked to join (although not risk free for certain).

I think the entire industry forgot about startup risk during COVID, and unfortunately it is coming rushing back now with the changing environment....

I definitely agree motivations for starting a company are mixed, and vary a lot from person to person. For the most successful companies I have seen, the motivations have largely collapsed to the ones I listed - with different mix of motivation by founder.

Finding the right word was definitely a struggle, but I think you need to lack alternatives in order to start a company. So "desperation" seemed to fit.

Open to other ideas/suggestions on a better word?

It depends on who gets vaccinated. For example, Israel focused first on the 20% of the population (largely elderly) who drive 95% of deaths.

Thanks for the fix! I also changed the title of my blog to remove "VR".

Too bad Twitter still doesnt have editable tweets :/

Thanks for the feedback. We will move away from calling it VR in the future and more about "spatial reality" or "spatial video" most likely

Yeah we changed names a week or two ago and the logo has not caught up yet :)

Hmm, I did not mean to abuse the term "VR". The original usage is much broader than just "put on a headset" immersive VR. For example, text-only MUDs & MUSHes

https://en.wikipedia.org/wiki/Virtual_reality

and

https://www.howtogeek.com/661428/text-based-vr-explore-the-p...

However, given how strongly people feel about the term, happy to use something different in the future to avoid ambiguity. E.g. "Virtual spaces" or "Spatial video".

Really the intention was to simulate aspects of the real world and in particular to virtualize human interactions in an online space.

Thanks for pointing this out.

Yorb was definitely one of the inspirations! Aidan is awesome and he attended one of the events we had.

Just looking for something space related :)

No headset needed. It works best on a laptop or desktop with keyboard right now versus mobile

Thanks! We view awkwardness as a good sign - it means we are hopefully capturing a part of real social interactions :)

Thanks! Let me know if any new requested spaces

Thanks! Let me know if you have any feedback.

Excited to share a side project I have been working on with my friends at Fluxon (Erad and AJ (@lasersox)) and Katsuya Noguchi (@katsuyan).

We got sick of all the Zoom happy hours, birthday parties, and company get togethers. In parallel, serendipity largely died - you could no longer run into a friend or co-worker or interesting new person given social distancing.

So, we put together Pluto - built on webRTC with MediaSoup and webGL with THREE.js and many other technologies.

Pluto is a browser-only VR space (no headsets needed) for spacial video chats and hangouts. You can move around in using first-person shooter-like controls. Sound decays with distance so different groups can break up and hang out.

Just drop in a link, and you can attend an event over video (no headset, no nothing). Create a virtual world in just one click!

Everything is represented in 3D which means people can randomly run into each other and serendipity can finally happen again.

We have built out a number of worlds using Blender to construct 3D models of : -Burning Man -The Maldives -Admont Abbey Library -A room in old Austria.

You can go to https://pluto.video to try it in your own VR in-browser world. It is best tried out with friends but can also be fun to explore the virtual world solo.

Come check it out and let us all know what you think!

It is still early / janky but we wanted to get it out there and get some feedback :)

We will be hanging out at an instance from 11am to 1:30pm PT at https://team.pluto.video.

Pluto was partially inspired in part by new social experiments on the web like Clubhouse, Gatherly, Rally, Rambly, Sonar, Teamflow, Yorb and others....

There are large amounts of the vaccine sitting on the sidelines right now.

See for example (scroll down and table shows % of doses used by state) https://www.nytimes.com/interactive/2020/us/covid-19-vaccine...

This is addressed in the post.

For example, New Jersey has 9 million people. Denmark 5 million. Why have other countries or states not achieved the same thing if size is the main driver?

What other companies are "index" companies? Spotify for music / streaming? Other?

I originally sent this out to the founders of companies I know. To your point it can apply more broadly.

Many larger company boards have already discussed this, but many startups have small or no boards to bring this up (hence the email I originally sent which turned into this post when founders suggested I publish more broadly)

Which of these do you think are farthest along?

Gene therapy seems to be working a little bit in ex-vivo approaches (you take cells out, modify them, and put them back in, largely in the immune system) but progress elsewhere is very limited.