Twitxy Saved

Morgan Lai | Ares.pro

Verified account ARES

@morganlai

building @arespro
@ Formulate Ventures @alliancedao @Stanford @MIT
NYC/SF morganlai.com Joined January 2012
5.7K Following 3.5K Followers 882 Posts

Latest posts 5

Pinned post

Morgan Lai | Ares.pro Verified account @morganlai
The future of consumer trading is social.
ARES @arespro Aug 19
Prediction markets are an information game. Most people play it blind.

Ares Mobile Beta is live: social trading for prediction markets, where every opinion has real money behind it. Powered by @Polymarket

First trade is free.
Drop your position card below → 25 winners get $25 announced on 8/27.
View quoted post on X
5 replies 0 reposts 21 likes 1.4K views View on X
Morgan Lai | Ares.pro Verified account @morganlai
There’s been a lot of discussion about how losing leaders like Jensen Huang would be disastrous for American technological leadership.

But last night in SF, I heard a different perspective from people working at the frontier of AI that I found particularly compelling.

Some researchers believe we could reach recursive self-improvement (RSI) within the next 12 months.

RSI is when AI becomes capable of improving the systems that make it intelligent. Better AI builds even better AI, potentially triggering an explosive acceleration in intelligence and scientific discovery.

Imagine decades of progress in medicine, physics, and engineering compressed into months or years.

But the same capabilities could also accelerate cyberattacks, biological weapons research, and military technology. This is why some researchers are discussing coordinated pauses and safeguards before RSI arrives.

Now consider Nvidia and Jensen Huang.

Jensen has consistently advocated selling advanced AI chips to China. One argument is that supplying China with American chips creates technological dependence, discouraging the development of domestic alternatives.

But look at TSMC and Taiwan.

China has historically depended on Taiwan’s advanced semiconductor manufacturing, yet that has never stopped Beijing from investing billions in domestic chip development or pursuing its strategic ambitions toward Taiwan.

Technological dependence doesn’t eliminate the pursuit of technological independence. Why would Nvidia be any different?

More importantly, we may be optimizing for the wrong variable.

What if the critical question isn’t who controls semiconductor supply chains in 10 years, but who reaches RSI first?

If RSI triggers the intelligence explosion some anticipate, the first country to achieve it could gain an extraordinary advantage in scientific discovery, cybersecurity, intelligence, and military capabilities.

In that scenario, selling advanced compute to a strategic competitor isn’t merely a commercial decision. It could fundamentally alter the geopolitical balance of power, particularly given China’s ambitions toward Taiwan.

Of course, RSI timelines remain speculative, and restricting chip exports carries real trade-offs. But the potential consequences of getting this wrong could be irreversible.

This brings me to the fundamental tension:

What’s best for Nvidia’s shareholders, what’s best for American technological leadership, and what’s best for American national security are three different optimization problems.

And as we approach potentially transformative advances in AI, we need to be much more deliberate about which one we’re optimizing for.
7 replies 1 reposts 18 likes 1.7K views View on X
Morgan Lai | Ares.pro Verified account @morganlai
I’ve always wondered whether humans would be averse to talking to AI agents on the phone, knowing they aren’t real. But after speaking with an AI agent to schedule an appointment today, I realized I actually much prefer it. The exchange of information is so much faster and more efficient, and you don’t have to manage anyone’s emotions or reactions. You can just get things done.
7 replies 0 reposts 14 likes 635 views View on X
Morgan Lai | Ares.pro Verified account @morganlai
Congrats @joshpxyne and @coframe! This is revolutionary for self-improving software development process. Proud to be an early investor backing this vision!
Josh Payne Verified account @joshpxyne 6d
The internet is dead. We’ve raised $35M total funding led by Khosla and Hanabi to bring it to life.

@coframe is a growth engine for your website/ads/more which tests, personalizes, & learns 24/7.

It’s working. Dropbox increased conversion by 158%. Replit, 410%. Cash App, 60%.
View quoted post on X
2 replies 1 reposts 15 likes 639 views View on X