Compiled by Waiwai Y.

Guest:

Imran Khan, Alliance DAO Founder

Qiao Wang, Alliance DAO Founder

Podcast Source:

Good Game Podcast

Original Title:

Trade War & The Markets | EP 74

Air Date: April 19, 2025

Highlights

Imran and Qiao sat down to discuss the trade war and the markets. No BS crypto insights for founders.

• Next year might be a good year for Bitcoin and risk assets.

• I am pessimistic about the future of the U.S. stocks.

• Qiao: I'm not currently taking any action on Bitcoin. My cost is basically zero, and the tax rate is about 30%. So I have to be very sure that Bitcoin will drop by 30% before I choose to sell, and I don't think it will.

• Imran: I sold all my stocks during the last rebound. I'm just planning to watch. I currently only hold Bitcoin, Solana, Fartcoin, and some cryptocurrencies. Fartcoin is performing quite well, and I think it's a safe haven.

• The Fed's dual mandate is to promote employment and control inflation, not to influence market prices.

• My biggest concern is not economic issues but potential military conflicts.

• Bitcoin has both Nasdaq's growth potential and gold's safe-haven characteristics to some extent, so it's expected to perform well.

• If this is indeed the case, I think under the Biden administration, there might be some regional or even global conflicts. Under the Trump administration, I'm not sure if he would particularly focus on the Taiwan issue.

• China is leading in many hardware manufacturing fields such as ships, automobiles, and solar energy, which are especially important during wartime. The U.S. has gradually lost this capability over the past few decades.

• If the purpose of tariffs is to bring manufacturing back to the U.S., then tariffs might not be the best strategy. A wiser approach is to invest in these industries, ensuring domestic companies are competitive in the global market, rather than harming the country's industries through tariffs, which is a zero-sum game.

• Once companies start using your AI products, they are less likely to switch to other products because the level of lock-in is deeper.

Eric Trump

Qiao:

Eric Trump was probably making billions of dollars through insider trading over the last few weeks. Eric Trump ended up selling at the bottom.

Imran:

Do you think that was just an illusion or pure stupidity? Maybe both. He was trying to run things for his dad, but then his dad kind of undermined him. He literally sold at the bottom.

Qiao:

It's insane. I saw this, and I think he used some of the cash to buy something. Did you see that? There was an announcement from a company, saying they'd buy a certain amount of tokens.

Imran:

Yeah, that was a while ago. It was pretty much a 3 and 3 with the Trump family and the powers that be.

Movement Labs

Imran:

Did you see the announcement? Did you see Movement Labs?

The 0 x balloon lever account has been on Rushi and Movement over the past six months, talking about insider trading, manipulating the token price, working with market makers that may not be professional. I don't know the full details. There was an event in China where people farming the airdrop got upset and disrupted the event Movement was hosting, which went viral. Influencers have been talking about how Rushi's not handling the token launch properly. Recently, Rushi announced he's taking a temporary leave of absence. That's where it stands. If you want to talk about that.

Qiao:

Has he sold his tokens?

Imran:

We don't know yet. OM Labs or Om lacks disclosures with market makers, founders, and exchanges. Many tokens seem great based on Twitter, but much of it could be manufactured.

Qiao:

How did OM go down 90% in a minute? Was it the market makers or someone else?

Imran:

Someone pointed the finger at Laser Digital, a trad firm in Southeast Asia, but they denied involvement. Hasib and others raised the issue that there's no transparency framework for token listings and market makers. This lack of transparency leads to manipulation of token prices, which may not reflect real market value. We're seeing a lot of that.

Market Discussion

Imran:

Let's talk about the market. We're probably bearish on the S&P 500. What are you holding right now?

Qiao:

I'm actually selling on stock rallies. Every time there is a fairly big bounce, I sell a little bit of my stocks, but I still hold the same stocks as before: Google, TSMC, PDD.

I'm really worried about a war. So I'm selling more of my TSMC than other stocks. My biggest holding is Bitcoin, but there's not much I can do with it. Bitcoin trades like a mix of S&P 500 and gold. Recently, Bitcoin's performance has been between NASDAQ and gold. I can't sell because my cost is zero, and the tax rate is 30%. So I have to be sure Bitcoin will go down by 30% to sell it, which I'm not. That's my holdings.

Imran:

I sold all my equities on the last pump. I'm just gonna wait it out. Right now, I'm holding Bitcoin, Solana, and some coins. Fartcoins has done well; it feels like the trenches of digital gold.

Capital flows for trenches went to smaller tokens. Remember Pepe? Similar market dynamics, bear market recovery, shot from 300 million to 3 or 4 billion. Similar dynamics now, just waiting to see where the market takes us.

Market Predictions

Imran: So before we were bullish on the S&P 500, we said that multiple times on our Pod. We were bullish on US equities before.

Qiao: I was long but also said it was very exclusive.

Imran: We both said the economy was great, so we were bullish. But this was before Trump shook things up.

Qiao: Before Liberation Day. Now we're coping. The reality sets in.

Imran: We're liberated. The market is liberated. Everyone's liberated.

Very bullish on crypto, even in our first few pods in January. But we felt exhausted. I mentioned in our first pod that although we're bullish, we're exhausted. Typically, that means you should sell, but we're not selling because we're crypto natives. Volatility. Now looking at it, we sold a little bit, right?

Qiao: I sold some. I don't know how much you sold.

Imran: I sold some and bought it all back. On this last big step, but I didn't sell a lot. Like 10%, not enough for me to say I sold whatever.

We're bullish short-term, bearish medium-term, bullish long-term. Is that how we want to think about it?

Qiao: Short-term, I'm bearish on US stocks. I don't have a strong view on Bitcoin because it's a mix of NASDAQ and gold. Next year feels like it'll be a good year for Bitcoin and risk assets. Trump says Powell will leave if asked. He adds he's not happy with him. If the market's gonna go down, Powell might leave. It may pump temporarily, but it's not good news. It's instability across the globe. Cool.

Regulation in Crypto

Qiao:

I mean, this is annoying because this was the problem that Messari set out to solve in 2017. Yes, right. And it's been 8 years. It is still an unsolved problem.

Like who owns what tokens, who is selling them? In traditional finance, with the SEC, whenever you're a major shareholder of a company and you're selling, you have to disclose it immediately, usually within a day or two. Like whenever Warren Buffett sells shares of a company where he's a pretty big holder, I always get an email of the SEC disclosure. There's a reason why these things exist.

Imran:

Agree. I think it's just the permissionless nature of the market where there isn't enough frameworks, regulatory guidelines. No one is really taking ownership of what this should look like. From what I see, they even talked about how we should have quarterly earnings processes. Every quarter, we should get on a call, go through a framework of what we've done, what we're about to do, and put up some forward guidance as far as where we think the market's gonna go and the company's gonna do.

Qiao:

The nice thing about crypto is you can do this daily, not even quarterly. Because everything is on-chain. If someone sells their insider shares, it's on-chain. All you have to do is link the address to the identity of the owner, and then you get automatic disclosure on a real-time basis. This is something that traditional finance cannot do, which is why they do quarterly reports and stuff like that.

Imran:

So I guess the question then becomes, who takes ownership of this? Is this something like a ByteDance would do? Is this something like a Jupiter would do? Or is it just more self-regulation?

Qiao:

This was the conversation we had at Messari seven years ago. Unfortunately, it turns out that this is not a problem that a startup can solve. Like Messari back then was like 5 people. This should be a problem that the big exchanges and other players at the top of the food chain should get together and solve. Unfortunately, that's not happening. It's definitely the responsibility of ByteDance, Coinbase, and others.

Imran:

I'd be curious to see how this would work. It would have to be some sort of alliance between exchanges, market makers, founders for this to come up. But I don't think this is gonna happen within the current regime.

Qiao:

I was thinking the other day, if this whole thing about tokens, if crypto failed, it would have been the biggest counterargument against the free market. In traditional finance, you could argue that it's not completely free market with the regulations, the SEC, and all that. There's a central sort of planner that dictates some rules. In crypto, we don't have those rules. It might just be the case that those rules are necessary and a completely free market may not be the ideal way of doing things. Maybe it should be 90% free, with 10% planned and regulated centrally.

Imran:

Yeah, I'd agree with that. I do think just a layer on top that slightly regulates how things get processed, IPO or ICO in our case.

Impact of Crypto's Speculative Nature

Imran:

It's probably because crypto generally has received some bad feedback due to its speculative nature, which has hurt our brand from a general perspective. High-quality talent will probably be less likely to join a certain tech sector because they see what's happening. If you see negative news on your timeline, are you really going to be excited about changing the world in the future? Probably not. But if you say stablecoin and frame it that way, you can say, well, we're crypto, but not really crypto—we're stablecoin.

Qiao:

It's not a perfect analogy, but it feels like the Web3 fork from crypto four years ago. Bitcoin maximalists have always tried to stay away from being associated with crypto. It's like Bitcoin and crypto are mutually exclusive. Now, stablecoins are distinct from crypto too.

Imran:

Distinct things. I think it's probably for the better. People in Silicon Valley, high-quality talent, may feel more comfortable joining this subsector as a way to differentiate from crypto. Oh, I'm not crypto. I'm stablecoin.

Qiao:

I don't know, whatever. Stablecoins are very easy for them to grasp. Even Bitcoin was not very easy to understand. But stablecoins are like fintech. It's basically Fintech 2.0. It's payment, remittance—all, all these things that traditional tech pretty much understands.

Trade War

Imran:

I think trade war, we're seeing a lot on the trade war side. Funny enough, Trump tweeted just today that we may have a deal with China, and then the markets pumped. Bitcoin is at almost 85k. I was listening to Powell yesterday in Chicago, and someone asked if the Fed would step into the stock market. Everyone laughed, and Powell said that the economy is doing great. The market is trying to digest all of this new information, and until that's done, it's hard to make any decisions.

Qiao:

Technically, the market is not part of the dual mandate of the Fed. The dual mandate is employment and inflation, not market prices.

Imran:

But you see how that is becoming part of the Fed's oversight, even though it's not mandated. The trade war, generally speaking, is really bad.

Qiao:

I think the trade war is worse than people expected in terms of its impact on businesses, consumers, and markets. Retail is still in the "buy the dip" mode. People think the Fed will save them, and US stocks are still very expensive. Even if you assume forward earnings remain the same, US stocks are trading at a PE of 24, which is high by historical standards.

Imran:

I think the market will wake up as soon as everything is finalized. People expect a V-shaped recovery once Trump announces the deal, but the psychological damage has been done. Consumers and businesses are cautious due to uncertainty, affecting the economy significantly.

Qiao:

The 10% tariffs are not small. If your business has a profit margin of 10% and costs go up by 10%, you're no longer profitable. Tariffs are at the highest level in history since the late 1920s and early 1930s.

Imran:

Toys in the US have seen an 80% price increase. Average vehicle prices will go up by 15%, adding $600 on average. Every household has to pay an additional $3,800 annually due to increased costs.

Trade War Strategy Discussion

Imran:

So what's the strategy here? I've read many perspectives, and some traders and macro experts suggest that the goal is to stop China's growth because it's becoming a world power and could surpass the US. The idea is to create a multipolar world where the US uses the dollar as a security blanket to encourage other countries to trade with the US but not China. They are trying to engage countries like Vietnam and Japan. If successful, this could reduce China's GDP and influence. What are your thoughts on that?

Qiao:

If that's the objective, it's inconsistent with execution. Trump has alienated not just China but also US allies like Europe, Japan, and South Korea. These countries are now talking with China about reducing tariffs. For example, Spain is exporting pork to China at reduced tariffs. Historically, China, South Korea, and Japan haven't been close, but now they are working together. The execution doesn't match the objective. It seems more like Trump is making decisions based on gut feelings without a real strategy, consistent with his past as a businessman.

Imran:

It's like negotiating out of a deal like high schoolers.

Qiao:

Yes, it's the high schooler art of negotiation. Overall, it's really bad and worse than people expect.

Impact of Trade War on Global Relations

Imran:

I think this trade war will have a worldwide impact, and it's going to be really bad for US stocks. People are now overweight in European equities, gold, and bitcoin for the first time in their lives.

Qiao:

My biggest concern isn't just the economy; it's the potential for military conflict.

Imran:

Are you thinking about China?

Qiao:

Yes, economic hardship increases the odds of military conflict, especially in the Taiwan Strait. Before the tariffs, I estimated a 25% chance of a military conflict between the US and China over Taiwan in the next three years. Now, I think it's 50%.

Imran:

Under the Biden administration, I could see regional or global conflicts. Under Trump, I'm not sure he cares much about Taiwan.

Qiao:

He cares about looking strong and not appearing cowardly. The probability of US involvement in a conflict with China over Taiwan has increased.

Imran:

If that's the case, there could be regional conflicts everywhere, not just Taiwan and China. It would be disastrous for global relations and trade. I hope we just fix this, enter a mild recession, and move on.

Impact of Trade War on Startups

Imran:

So how does this affect startups locally? I mean, like I could tell you, we may not get founders from China and India as an example, but what are your thoughts?

Qiao:

Don't you think that venture capital is gonna slow down if the economy slows down? Don't you think that will affect venture markets as well? The venture markets tend to lag the public markets by a couple of quarters.

Imran:

I think depending on where we are with AI. Cuz AI is pretty hot right now. And, you know, like I just did a quick search where it scraped all the data from all the YC startups over the past two batches. And I asked it like, what's the average valuation post demo day? And the average valuation was about 22 million for an AI startup. To me, that's pretty high, right? You know, you wanna be in the 15 million dollar range, right? That's kind of the sweet spot. The 22 million is like without any product or even an MVP is pretty high to me. And I feel like it's just gonna go higher at this point.

So I do think we're getting to that point where things could pop in the AI side, especially given what's happening with like, maybe not now, maybe next 6 months, a year, but like, you know, what OpenAI is doing with Winsurf is a perfect example of like, no one is safe in any product that they build unless they have some sort of network.

OpenAI Tried to Buy Windsurf

Imran:

I saw yesterday I posted in our chat about how OpenAI tried to acquire Windsurf. Sorry, I'm thinking Windsor for that.

Qiao:

Was it like $3 billion? And they apparently approached Cursor two times to acquire them and they said no. So then they went out and acquired Oneserve for $3 billion. There's a lot at play, right? You could think of OpenAI as kind of like a layer one, and then you have these apps that are being built on top. Whoever has the most apps gives them the best models, which attracts more people using the product that helps them build more apps. Is that the way to think about it?

Well, there's two things that I thought of. One is that so far, OpenAI has not produced the best in class coding models yet. They've been trying to, but Anthropic is probably the leading model when it comes to coding. Cursor and probably Windsurf are both using Anthropic. So from OpenAI's point of view, they want to build the best in class coding model, but that requires a lot of data.

Where do you get the data? You can buy the data from companies that specialize in data labeling, like Scale AI, for actual coding data. Another way is to get it directly from Cursor or Windsurf. When you use Cursor and Windsurf to code, you effectively give data to them by prompting, asking for code suggestions, and accepting the code. This is like reinforcement learning with human feedback. This data is very valuable, and that's why OpenAI wanted this data. They wanted to buy Cursor, and when Cursor refused, they got Oneserve. That's the first thing I thought of.

The second thing is that those big AI labs are seeing where the market is going, where the eyeballs and users are. They are probably worried that the underlying models will get commoditized over time, similar to crypto. You have the layer ones that are likely to get commoditized, and it's the app layer that's going to capture a lot of value. By acquiring an app with many users like Windsurf or Cursor, they can solidify their position on the app layer. ChatGPT is an app built on top of various OpenAI models. But they want more, like coding apps and maybe something else in the future. That's where I think they're going with this acquisition.

OpenAI as a Super App

Imran:

To me, this seems more analogous to a super app where, and I'm sure you've seen the announcement around OpenAI exploring ideas around social as well. It's almost like a one-stop shop to experience AI in different modalities, whether it's coding, photo and video generation to create content and story, or just general intelligence for day-to-day stuff that you would use Google for, as an example. It's interesting because you could argue that for every category OpenAI introduces a product for, it cannibalizes hundreds of startups building for that specific sector. Even within the lines we have, we carved out a very specific niche around leveraging crypto to build apps, and we think there's a unique mode there because of the crypto aspect. But you could argue a lot of this stuff could be cannibalized by the larger players.

Qiao:

Actually, Vinod Khosla said recently that half of the YC badges are basically dead every time a new foundational model comes out. Not dead, of course, they can always pivot, but their current idea is no longer viable whenever a new foundation model emerges.

Investment in AI Companies

Imran:

Another interesting point, I don't know if you saw that graph, but there's a graph of like six VCs—Index Ventures, Andreessen Horowitz, Sequoia Capital, maybe Lightspeed, and others like First Round. It showed how all the major AI companies are overlaid over it and who made the investments. Most VCs have investments in two to three of these startups. Vinod Khosla had only one investment, which was in OpenAI in the early stages, and that was it.

Funny enough, I was talking to Founders Fund yesterday. They've only made one major investment in the AI space, which was primarily OpenAI, with one other smaller investment in a very specific niche AI company. When I was talking to them, they said Peter Thiel does not want to over-invest in AI. He thinks it's all going to get cannibalized like the dot-com bubble. He told the team not to focus too much on AI because it reminds him a lot of the dot-com bubble.

Qiao:

I remember he said that maybe half a year ago on the All-In podcast. Yes, I watched that interview. He drew parallels between AI and the dot-com bubble.

Imran:

And he reiterated that when they launched a new fund recently. He explained again why they should not be investing heavily in AI right now.

AI vs Crypto Startups

Imran:

In a way, it's too consensus. There's just too much competition. If you thought crypto was competitive at this moment, you should see what's happening in AI. We have a few AI startups in our current batch, and for every startup, there are probably 25 to 50 competitors.

Qiao:

I think right now, the level of competition and crowdedness in AI is probably one or two orders of magnitude higher than for crypto startups.

Imran:

I feel like crypto's becoming non-consensus again to a certain degree.

Qiao:

It's actually gotten more and more non-consensus over the last two years.

Imran:

To the point where they're trying to reframe crypto in a completely different aspect now. If you look at Zach, I think Zach is a founder of Bridge.xyz, and a week ago he tweeted that the next big platform should be stablecoins after cloud, mobile, and the internet. He completely cut out crypto from that conversation as the next paradigm shift.

Imran:

And so I thought that was also a very interesting narrative they’re trying to run over there.

Exciting Startups in AI and Crypto

Imran:

Any startups that you're excited about? We kind of see the future a quarter ahead to a certain degree with the types of founders that come into our program. So if you had to call out a few startups, what startups are you most excited about?

Qiao:

We have many stablecoin startups in the current batch. A third of the batch is focused on stablecoins, targeting specific regions or use cases like remittance or yield. We also have several AI startups focusing on specific verticals, such as a cursor for scientific research, video ads using AI video models, and coding for crypto apps. They're all very interesting with very good teams.

Poof.new

Imran:

The startups I'm really excited about working with are the 3 to 4 that I'm most excited about. One is the one you mentioned, Poof.new, which is the vibe coding platform for crypto. I think they have a unique edge there. The team is very strong; they came from Phantom and Coinbase and built a similar product before on Flow 2-3 years ago. They have the domain expertise to build something great here, in my opinion.

Slop.club

Imran:

Bit more experimental. I know how you feel about it, but Slop.club is a video generation platform. It uses advanced models like Clean and Vo from Google, allowing people to create and remix content. It's cool to see people upload pictures and create fun video models around them. There's some interesting consumer behavior happening. OpenAI's wedge in the market is image and video creation, and Slop.club is analogous to 4chan but for video models.

Qiao:

The core insight is that video models enable new behavior that wasn't possible before, reducing latency in video creation and remixing. Previously, creating a video took a long time, but now you can go from a prompt to a video in seconds or minutes. The ability to remix on the spot is similar, allowing instant creativity compared to longer processes before. This is the core insight behind the product.

01 Exchange

Imran:

Last one is 01.exchange, which involves Photon and Bullex. You might have seen tweets today that Bullex founders aren't very responsive. They've made a ton of money, and founders' incentives in crypto can be challenging. Even if you're a billionaire or have hundreds of millions, do you want to build a groundbreaking product? The best founders overcome these challenges. The Pump team is a perfect example; they've earned tons of revenue and continue innovating, like announcing live streaming recently. It's a unique battle for founders.

Bullex kind of took the pedal off the gas. Photon, I'm not sure where they are from a product perspective. But 01.exchange, the team is hungry and working tirelessly. They're building unique ways to upgrade products for traders. These are some of the exciting products.

Impact of Military Conflict on AI Progress

Qiao:

I was thinking how potential military conflict could cause supply chain disruption in semiconductors, which would ultimately affect progress in AI and robotics. Robotics would be affected to a large extent because it's more hardware-dependent.

And a lot of the hardware comes from China too, right? A lot of the advanced chips, though not the cutting-edge 3 nanometer ones, come from China. Many other electronic parts are also manufactured in China.

China's Dominance in Robotics

Imran:

I was reading some stats on China's dominance in robotics. Although I haven't looked into humanoid robots, China manufactures a ton of industrial robots globally, holding a 51% share of global sales. In comparison, the US has only 5%.

If humanoid robots were to scale up in production and sales, China would primarily own that market, not the US.

Qiao:

I feel robotics will follow the same path as digital AI, like Agentic AI. The US will innovate and be the first mover, creating new things, while China will make them cheaper and more efficient.

Imran: Did you watch the All-In podcast with Travis Kalanick, the founder of Uber?

He talked about competing with Chinese competitors. He was annoyed that every time he released a new feature, Chinese companies would copy it within a day. The speed of copying was so fast that he developed respect for these startups.

Qiao:

It's different now. Even with BYD cars, they have features like parallel parking and advanced suspensions. It's impressive, but you can't buy them in the US.

I spoke to someone knowledgeable about Chinese EVs. They look luxurious and are cheaper than Tesla, with good batteries. However, they're not safe; there are stories of cars blowing up due to battery fires. These incidents are quickly censored.

Tesla is way ahead in Full Self-Driving (FSD) technology compared to Chinese EVs.

China's Advancements in Solar Energy

Imran:

What other products do you feel like China's ahead in compared to the US?

Qiao:

Solar energy, definitely.

Imran:

I think they manufacture about 80% of solar panels globally.

Qiao:

Something like that. All this progress happened in the last 10 years. China had nothing a decade ago and then invested heavily in these industries. I feel there's a strategic reason for China's investment in hardware manufacturing, as these industries become crucial during wars. China is now No. 1 in manufacturing ships, cars, solar energy, and other hardware. These things start to matter during conflicts, and the US has lost that capacity over the last few decades.

Imran:

I don't think the US had it to begin with. Even reading through the "Chip War" book, the US was lagging in chip production yields.

Qiao:

This goes back to the tariff war. If the goal of tariffs is to bring manufacturing back to the US, then tariffs aren't the best strategy. The smarter approach is to invest in those industries, ensuring your own companies are competitive globally rather than inflicting damage on your industry and country. It's a negative-sum game.

Imran:

It's negative-sum. You need to compete harder and faster, reinvest in US manufacturing, but keep prices down to maintain consumer confidence globally. People say Trump is playing 4D chess, but...

Benefits of Going Public

Imran:

I was reading about Snapchat's founder, Evan, who recently talked about the benefits of going public. He said that IPO made them much more operationally efficient, both from a back-office perspective and user acquisition perspective.

It helped him run a tighter ship. There are positives to going public, as it forces founders to manage capital more appropriately and focus on user acquisition. When releasing new products, it encourages founders to think long-term, not just from a 6-month perspective but 5 to 10 years out. This layer of operational excellence is important when going public.

Zora Launch & "Base is for everyone" Token

Imran:

Did you see the Zora launch yesterday and the Base "Everything is for Everyone" token?

The token was called "Base is for Everyone." It reached a $50 million market cap. I thought about buying it but decided to wait since it's a semi-bear market. It crashed to $1.5 million, and I thought it was a good buy then. People were upset because a sniper bought 25% of the token and dumped it, causing the crash. This highlights issues with bonding curves and snipers, which is a problem across platforms.

The narrative that Zora and Base are trying to create is about tokenizing all forms of content, allowing content creators to earn revenue based on their on-chain content. However, 75% of crypto Twitter criticized Zora and Base for this rug pull, and many are just engagement farmers.

Qiao:

They call it "Content Coin." What do you think of it?

Imran:

I was going to ask you that. It might be a narrative for normies, but I don't think I'm the right user for it. I've been using Zora for the past six months.

Qiao:

And what do you think?

Imran:

I like the flow. Would you use Zora or buy content coins?

Qiao:

It's just a new term for meme coins. I don't think this term will catch on because there's so much inertia with the meme coin terminology.

Imran:

Everyone knows meme coins. I could tell my mom about them, and she'd know what they are. It's become analogous to something like Bitcoin. But I will say Jesse and Base are fighting hard for their founders, which I commend.

What's New in AI?

Imran:

What's new on the AI side? We talked about acquisitions, but what products are you thinking about where founders wouldn't worry about disruption from OpenAI or other larger AI companies?

Qiao:

I think enterprise use cases are safer than consumer ones. Enterprises are internally more resistant to change, so once they use your AI product, it's harder for them to switch.

Imran:

I agree. Enterprises are resistant to change because products are often sold in bundles. For example, Microsoft products like Office 365 and Teams lock enterprises in.

Qiao:

Some people use all Microsoft products.

Imran:

Exactly. Microsoft Teams was given for free to Office 365 subscribers, which impacted Slack's growth. Enterprises have a better chance of converting customers to new products because they're already locked into subscriptions.

For example, Canva launched a product for education, targeting schools and universities as enterprise customers, allowing any kid to build an app. This impacts competitors trying to reach these markets.

But here's the counter-argument: Zoom became successful because consumers used it during Covid, which then bubbled up to enterprise adoption. The same happened with iPhones replacing Blackberry in enterprises.

Qiao:

You're not arguing against my view. Enterprises are harder to disrupt by OpenAI compared to consumer products.

Imran:

OpenAI could build enterprise products because they already have consumer support. They've become synonymous with AI, making it easier to win adjacent products.

Microsoft and OpenAI are competing because OpenAI moved workloads from Azure to Oracle, counterbalancing Microsoft's competition. OpenAI is battling on too many fronts, which can lead to failure.

Qiao:

Back to the original question: AI apps that are harder for OpenAI to disrupt can choose from various foundational models. For instance, Cursor and Windserve use Anthropic for better coding models than OpenAI.

Imran:

It's network effects, essentially.

Qiao:

It's the ability to choose different models. You only get that leverage if you have distribution.

Imran:

You get distribution by building something useful first.

Qiao:

It's hard to predict which verticals are resistant to OpenAI. It depends on how fast an app can acquire users and what OpenAI wants to pursue.

OpenAI o3

Qiao:

So OpenAI released o3. The benchmarks look pretty good for them.

Imran:

People call it AGI. I hate these examples, but have you seen those tweets? They're like, "Oh my God, it's so good."

Qiao:

And then someone posted a simple math question it couldn't solve.

But o3 ranks pretty high. It ranks better than Gemini 2.5. So O3 is now the best model in its category, like a long context reasoning model.

Competitive Nature of AI Models

Qiao:

The advantage that the app layer has is the ability to adapt to the latest models, whereas OpenAI can only use their own model.

Imran:

It's almost as if models come secondary now because distribution is key.

Qiao:

Right. It's much easier to switch models than chains. AI models are less defensible and more commoditized than blockchains.

Imran:

A perfect example is Zuck releasing Llama 3 and Llama 4. Even that side is becoming so competitive. No one wants to use Llama 4.

Qiao:

It's so behind.

Imran:

Even researchers that contributed to Llama 2 and 3 have left for OpenAI. If you look at LinkedIn, it's funny.

Qiao:

I saw that. He said he worked on Llama 2 and 3, but not 4.

Imran:

Everything is so competitive. If you don't have money or the best people, you have no edge. Open router and Freedom GBT aggregating models is an interesting play.

I pitched Freedom GBT to people. No one knows anything outside OpenAI. Kids aged 12 to 18 haven't heard of anything else. I explained the concept, and they finally got it.

Qiao:

I use AI by asking questions to multiple models like ChatGPT and Gemini, sometimes DeepSeek Cloud, to triangulate answers.

Imran:

Certain models have key advantages. Google Gemini scripts YouTube videos and proprietary content, pulling sources and providing nuanced information. It's an interesting wedge.

AI Memory Feature

Imran:

We discussed memory. OpenAI did a big upgrade on memory. I'm curious about your thoughts.

Qiao:

I think there's some moat, but it's not that significant. Memory is useful, but I'm not sure what "full memory" means. I've asked ChatGPT before about my MBTI, and it gave a good answer. I thought memory was already accessible.

Imran:

From what I read, they scaled up memory to remember more about you across conversations. There's a memory button in OpenAI, and I reset it sometimes. It's good that it has context about me for tailored answers. It feels like it knows me, which is weird, but I appreciate it. Some people might feel like they're having relationships with it.

Qiao:

It's like Iron Man and Jarvis.

Imran:

I don't feel it's sentient, but it tailors conversations with glimpses of human emotional links. It empathizes with you, and that's weird. There's some moat in that because the more connected you are, the harder it is to switch.