Compiled by Waiwai Y.

Gain actionable insights on how tariffs, macroeconomics, and AI are influencing the crypto market and blockchain ecosystems.

Guest: Imran Khan, Co-founder of Alliance DAO; Qiao Wang, Co-founder of Alliance DAO

Podcast Source: Good Game Podcast

Original Title: How Long Will This Bear Market Last? | EP 72

Date: March 19, 2025

What’s next for the market cycle? In this episode, we explore whether the recent peak marked the top or if there’s still room for growth in the months ahead.

Highlights of Key Insights

• At this level, I don't know if we've seen the mid cycle bottom yet, but this is a really good level to do long.

• The cycle isn’t over, and that likely wasn’t the top. We’re probably mid-cycle, with 6 to 18 months remaining.

• The market just needed an excuse for a correction, and that excuse turned out to be tariff policies.

• So this breakdown was actually a good thing, because it gives more breathing room.

• In general everyone on Twitter is being affected by everyone else's sentiment. And I think basically everyone is wrong. Everyone was all the bears were wrong. All the bears.

• Institutions prefer tokenized t-bills over stablecoins.

• Like the idea of tokenizing early-stage ideas and products. Essentially, this falls under the broader category of tokenizing startups, equity, or even entire companies.

• Web2 giants like Robinhood and Kalshi might gradually squeeze out crypto startups, making it harder for them to survive.

State of the Market?

Imran：

This guy is named Ki Young Ju. His most recent tweet said, Bitcoin bull cycle is over. Expecting 6 to 12 months of bearish or sideways price action.

Qiao：

What's the reasoning? Like what are these people's reasoning?

Imran：

I'll read it through. Every on chain metric signals a bear market. With fresh liquidity drying up, new whales are selling bitcoin at lower prices.

The alert applies PCA to anchor indicators like MVRV, SOPR and NUPL to compute a 365 day moving average. The signal identifies inflection points where the trend of the one you're moving average changes.

Qiao：

I'm really skeptical of launching data this cycle, cause a lot of the big one is now with ETFs.

Imran：

Some specifically called the bear market because of more of the macro, which is the tariffs. There's an impact on tariffs. Obviously, inflation hasn't come down yet. He also mentioned the macro changes on AI stuff.

Tariffs & the Economy

Imran：What are the things that we're saying that keeps us bullish?

Qiao：

The economy looks good. From the election to now, or basically over the last few months, earnings forecast, corporate earnings forecast have gone up, not down.

The high yield spread is near historic lows. So like high yield spread is generally a sign of credit risk in in the market in the economy. Like when there is a recession, credit risk goes up cuz companies can't pay their debt anymore.

It's near historical, although it has gone up a little bit because of the tariffs. The fear around Terrace inflation is down actually. Inflation has gone down in the short term.

Which is the opposite of what people would expect from the tariffs cuz people think tariffs are inflationary, but that's not happening. The inflation is actually going down, not up. And the unemployment rate is also near the lows. So I actually think that there's almost no cracking the economy. Of course, the risk is higher then three months ago, right before the whole tariffs thing started.

Like I'm not saying tariffs are nothing burger. Of course, it'll impact how people spend, especially how companies spend. If there's uncertainty around how much they're gonna pay to import some goods, then they're less likely to spend, right? They're less likely to conduct their normal business. Uncertainty is actually a big risk. It's a big problem for companies and individuals, but none of that isn't the data yet. So I think the economy is actually doing pretty well. And given that and the fact that bitcoin is now a macro asset class, I think over the next 6 to 18 months period, bitcoin is actually gonna do pretty well.

Imran：

I was reading through like the actual impacts of tariffs specifically in the US. And, you know, automakers like BMW, they're taking steps to absorb some of the costs temporarily. Awesome. And so they're gonna essentially offer price Protection for consumers in the US.

I'm not sure if this is gonna be long term, but in the at least short to medium term, there's gonna be some price protections, at least within the auto space. And I would assume this is probably gonna happen in other sectors as well. And so I do think even if tariffs were to come, it's gonna take some time before consumers actually.

Qiao：

If the longer this thing drags, yeah, the more risky it is, like inflation can actually go up and become sticky if Trump doesn't shut up about terrorists for much longer. And there will be cracks in the economy. What I'm saying is it's not there yet. And did you notice that Trump has shut up about tariffs over the last few days? I feel like it's near his max pain, his own personal max pain.

Imran：

This typically occurs during the first 100 days of a presidential term. During this period, presidents often make significant economic adjustments or redefine how policies are implemented. Afterward, policies tend to stabilize. However, we still need to keep an eye on the situation. I heard that he plans to announce new policies on April 2. I saw his tweet this morning, suggesting that April 2 might be the deadline, and he is planning to impose universal tariffs on 20% of the world's countries.

Qiao：

And this goes back to my point about tariffs. The worst thing about tariffs is uncertainty. It's actually not the tariffs themselves. The longer this whole thing is uncertain, the more actually it'll impact the economy. But if Trump just tell us, hey, this is what we're gonna do, then there is no authority, like business can adjust. And that will be, we will be in a much better situation.

Imran：

I guess Trump's other mandate was to bring manufacturing back to the US and, you know, US goods should Trump goods that are coming from outside of the US. And Anthon argued that US manufacturing is actually very weak compared to, let's say, Shenzhen or China.

Qiao：

But this is a longer thing. There's not a, it's not a medium term economic cycles kind of thing, it takes years to bring manufacturing back and to see the effect.

Imran：

But the argument there is because it's gonna take so long, consumers will not see the impact right away because. So we'll have to see how that will impact consumer pricing.

Crypto Sentiment Divergence

Qiao：

Here's the other thing I noticed. There is a huge discrepancy between crypto-native sentiment and the normal traditional finance institutional sentiment. I've never seen this much bearishness from the kryptonatives since FTX.

And then among the institutions, they're extremely bullish. The banks are opening up, they're starting to relax how much their customers can invest in Bitcoin ETFs.

Imran：

I'm gonna give you some data. Behind what you're just saying, since Trump's election, we've added $50 billion in stablecoin. Second, we went from roughly less than $2 billion to 4.1 billion in tokenized treasuries. Anything with RWAs is up in terms of AUM.

Institutions prefer tokenized t-bills over stablecoins. And the specific reason behind this is because stablecoins have some sort of risk involved with, whereas with t-bills, there's more of a secure off ramp with institutions that are accustoming.

Then you probably saw the other bullish aspects of RWAs and TradFi were the announcement of Ethena and securitize with their layer one called Converge. And then Ondo chain also. So there's a lot of bullish news that's coming out around the RWA in TradFi world.

Qiao：

I think in general everyone on Twitter is being affected by everyone else's sentiment. And I think basically everyone is wrong. Everyone was all the bears were wrong. All the bears.

Imran：

I think crypto Twitter has just become an echo chamber, so much so that it's creating a negative cycle for everyone that's involved. And that I think what people should do is disengage from crypto Twitter more and just focus on what's happening on chain and all the other aspects of crypto in terms of growth. And I feel like people are doing that much less right now.

"Where are we and where are we headed?"

Imran：

Thoughts on crypto. Where are we headed?

Qiao： I mean we sent the alarm at the top two months ago.

Imran：

We're just exhausted. I think they were mentally exhausted.

Qiao：

But where I got wrong was I thought we were near the top of the cycle. In hindsight, I don't think the cycle is over. I don't think that was the top. I think we're probably actually mid cycle and there's probably maybe 6 to 18 months more to go.

Imran：

So this breakdown was actually a good thing, because it gives more breathing room.

Qiao：

This whole breakdown was due to sentiment around tarrifs. Bitcoin followed the stocks and stocks in early January was trading at historic high valuations. In terms of PE ratio, US stock were trading at the same level as mid 2021, before Federal Reserve Chair Jerome Powell started raising interest rates, and even approaching the levels seen before the 2007 global financial crisis. So, U.S. Stock valuations were extremely high.

The market just needed an excuse for a correction, and that excuse turned out to be tariff policies. Now, the market has adjusted—U.S. stocks have dropped at least 10%, and the Nasdaq fell by 13%. I think this correction was necessary. While valuations are still high, they’re now more reasonable.

However, in terms of the current economic situation, there aren’t any glaring issues. If Trump could stop talking about tariffs so frequently, I think the economy might improve. But I feel like he’s approaching his “personal limit,” or the point of maximum pressure.

Imran：

They always claim they don’t care if the stock market enters a recession or if the country falls into an economic downturn, but I don’t believe that.

Qiao：

I'm 80% confident that the cycle top is not in. I think at this level, I don't know if we've seen the mid cycle bottom yet, but this is a really good level to do long.

There might be some emerging sectors, and some narratives could become important. But I don’t think there’s a clear narrative at the moment. In the future, we might see some key market narratives emerge.

Imran：

Previously, people lost a lot of money on meme-related tokens, and now they’re starting to focus on tokens that can actually generate revenue.

Qiao：

That’s what people are discussing online, but the market doesn’t always operate according to this logic. However, we can keep an eye on it. For now, I haven’t seen anything particularly compelling outside of Bitcoin. Ethereum is approaching historically oversold levels, but I can’t find a strong enough reason to hold it. It’s hard to buy into it, but at the same time, its fundamentals are actually decent.

I still believe Bitcoin is the asset I’m most optimistic about. Recently, I’ve also been buying some stocks, including Google, TSMC, Tesla, and Pinduoduo (PDD)—companies we’ve discussed over the past two years. Except for Tesla, I’ve finally increased my position in Tesla recently. Back in November, I tested Tesla’s FSD (Full Self-Driving) system, and it’s truly a game-changing product. So, I’ve been waiting for a price correction. After Tesla dropped 50% from its peak, I finally bought in. b

Ethena and Ondo Launching Their Own Layer 1

Imran：

What do you think about Ethena and Ondo launching their own Layer 1 blockchains? I find it quite interesting, especially with their permissioned validators coming from traditional financial institutions. This, to some extent, undermines Ethereum's moat in the DeFi space. After all, one of Ethereum's key advantages is its massive asset pool accumulated over the years.

Recently, Standard Chartered published a research report mentioning that Base has "extracted" around $50 billion in value from Ethereum's ecosystem, right? So, in my view, a lot of value is now flowing from Ethereum's Layer 2 and Layer 1 to other chains.

Qiao：

What do you think about this? I don’t want to argue about Ethereum anymore—this is just the current reality.

Imran：

It’s clear that this isn’t good news for Ethereum. I believe this trend will continue in the future. RWA (Real-World Assets) might opt to build their own blockchains, or so-called “permissioned chains,” because these chains can offer more functionality.

For instance, they can choose to “rollback” data on the chain, and they have greater control over the ledger.

Regulation and Counterparty Risk

Qiao：But I don’t quite understand why they’re doing this in a permissioned way. What’s the point? What do they gain from it?

Imran：

One reason is greater control over the ledger. For example, if there’s a hacking incident originating from a country like North Korea, it’s almost impossible to address this on a permissionless blockchain. But on a permissioned chain, they can take measures to mitigate losses by controlling the validators.

Qiao：

But who are they trying to attract with this permissioned chain? For instance, if you want to buy their products today, like money market funds or similar offerings, you can just open a brokerage account.

The significance of tokenized funds on Ethereum is that it allows people who can’t open brokerage accounts to still purchase these products via Ethereum. So, with the creation of a new permissioned chain, what problem are they actually solving?

Imran：

Perhaps it’s for global access. Knowing who the counterparties are can make the entire system more regulatory-friendly. Since they understand the roles of all participants in the value chain, there’s no need to worry about counterparty risk, the anonymity of the other party in a transaction, or security issues like hacking. Do you think this could be one of their considerations?

Qiao：

But does this bring them more business? Every time a new chain is created, there’s a huge challenge in user acquisition. You have to figure out how to attract users.

Imran：

Traditional financial institutions already have their own distribution channels. Perhaps these traditional financial fund companies have some sort of partnership with these two Layer 1 projects, but I’m not sure how deep that partnership goes. It could just be co-marketing, and we don’t know the exact scope of the collaboration.

What I mean is, I’m just trying to think from their perspective to see if there’s any real value here, or if this is just “smoke and mirrors”—something that looks complex but lacks substance.

If they do have distribution channels and are operating on a permissioned chain, they might simultaneously address counterparty risk and hacking issues while offering a globalized product. That’s the argument in favor of this model.

But the counterargument is that Ethereum is permissionless—anyone can build applications on it, and anyone can use it. All you need is an internet connection to access the global market without needing to go through KYC.

Qiao：

I was surprised by Standard Chartered’s mention of Layer 2. They seem to believe that Layer 2 can retain a significant amount of sequencing fees instead of passing all those fees to Layer 1.

Tokenizing Coin on Base

Imran：

Have you seen their announcement about tokenizing assets on Base?

Qiao：

Are they really planning to do this?

Imran: I think they’ve already started, or at least they’re testing it.

Qiao：

I believe that on a global scale, Coinbase and Base might be among the best-positioned companies to succeed in tokenizing stocks.

Not only do they have strong distribution channels, but they also have extensive experience in handling traditional finance, right? They are one of the primary custodians for many large asset management funds. Moreover, as a publicly listed company, they bring an added layer of credibility. And being based in the U.S., under strict regulatory oversight, is undoubtedly a huge advantage.

Imran：

So, in the RWA space, they might have a stronger edge than Ethena, Ondo, or any other competitors.

Robinhood & Kalshi

Qiao:

Another company that might be very suitable for tokenizing stocks is Robinhood.

Imran: If you look at their recent announcements, they’ve mentioned not just tokenized stocks but also real estate.

In fact, their founder, Vlad, mentioned that he envisions building a platform where users can easily buy and sell real estate with just one click.

They’ve also recently partnered with Kalshi. I remember we discussed this before. Kalshi has already signed an agreement with the U.S. Commodity Futures Trading Commission (CFTC), allowing them to offer prediction market services in the U.S. These services include specific contracts resembling prediction markets. Kalshi is currently the only federally authorized company that can legally operate such services across the U.S.

Robinhood has just established a partnership with Kalshi, enabling all Robinhood users to access comprehensive prediction market features. This raises some concerns.

Web2 giants like Robinhood and Kalshi might gradually squeeze out crypto startups, making it harder for them to survive. This is an issue worth paying attention to, as it could significantly impact entrepreneurs in the crypto space—at least in the U.S. However, I think there’s still significant growth potential for crypto globally. For example, Polymarket is still performing quite well in international markets. What’s your take on this?

Qiao:

I think they’re not our allies—they’re more like external threats.

Imran:

There are indeed many startups and traditional companies that pose threats to entrepreneurs in our ecosystem. So, the only way we can respond to these challenges is by moving quickly and executing efficiently.

State of AI

Manus AI

Imran:

I recently noticed that half of my feed is about AI startups. In fact, we mentioned something similar in our last podcast. Among the startups we’re following, four are fully focused on AI. A lot of interesting things are happening in the AI space lately. Have you used Manus?

Qiao:

I don’t have access.

Imran:

I used Manus and Operator to complete some simple tasks. Operator is a tool launched by ChatGPT or OpenAI. I tried two tasks, like fetching the latest batch of startups from YC’s (Y Combinator) website and filtering out the crypto startups that are already live.

I did the same task with both Manus and Operator. Manus took about 4 minutes to complete. In 3 to 4 minutes, Manus provided a full result, including the names of the startups, the founders, and their specific business focus. It essentially scraped all the relevant information.

Operator, on the other hand, got stuck. After running for two minutes, it returned: “I checked the latest batch of startups on YC’s website and only found one company named Lero, which focuses on trainable deep agents and artificial intelligence.” It also added that there were no crypto-related startups in this batch. Manus’s performance was impressive in comparison.

Qiao:

Do these tools run the tasks on your computer, or do they rely on the cloud?

Imran:

I believe they run in the cloud rather than directly on my computer. For example, Operator shows a desktop interface, simulates mouse movements and clicks, searches YC’s information, browses the site, and scrapes content.

This also reminds me of how many Chinese startups seem to be ahead of the U.S. in certain areas. At least from discussions on Twitter, external observers generally believe that Chinese startups have surpassed the U.S. in some fields.

Qiao:

Over the past month, I’ve come to a similar conclusion. I’m not sure if China is outright leading the U.S., but at the very least, they’re on par. For example, DeepSeek is a standout case. Everyone knows DeepSeek’s performance is almost comparable to top-tier U.S. models, but its scale is one to two orders of magnitude smaller. This means it can run efficiently on local devices.

China vs. US AI Video Models

Qiao:

There’s also an interesting trend. In our incubator, we currently have several video-based AI startups. They are using three to four video models, and only one comes from the U.S., while the other three are from China. Reportedly, the Chinese models not only perform better but are also cheaper and of higher product quality.

That said, AI’s applications are incredibly broad. These are just scattered data points. For instance, you could argue that Tesla’s autonomous driving technology is a form of “physical AI,” right? Tesla’s FSD (Full Self-Driving) tech is globally leading. So, it’s hard to definitively say which country has the upper hand in AI. But I think China has at least reached the same level as the U.S. for now.

Imran:

I feel the same. Especially after the Biden administration restricted the export of high-end video chips to China, China has clearly increased its investment in domestic chip manufacturing, particularly in collaboration with SMIC (Semiconductor Manufacturing International Corporation). Have you noticed this?

Qiao:

But their chip technology is still generations behind.

Imran:

Even so, companies like DeepSeek are already leveraging existing chips and software to develop products that can compete with U.S. startups in the media space. While they haven’t completely caught up, the gap is narrowing. I think maybe after one or two more technological iterations, Chinese products will reach the same level as U.S. ones. It’s foreseeable that they’ll continue to increase investment in SMIC.

Qiao:

In fact, the chip export restrictions might force China to innovate in terms of model scale.

Imran:

In the long term, this might actually be an advantage for China, making them more resilient and innovative in their competition with the U.S.

Qiao:

Models like DeepSeek are open source.

Imran:

That’s interesting because I haven’t heard much about Llama 3 recently.

Qiao:

There’s a noteworthy trend here. First, the technological competition between these two superpowers has essentially become an “arms race,” with both sides rapidly catching up to each other.

Localized LLMs & Privacy

Qiao:

Our team recently discussed an interesting trend: localized large language models (LLMs).

With advancements in technology, the size of LLMs is gradually shrinking without compromising performance. This means they could eventually run directly on local devices, such as personal computers or even smartphones. This opens up possibilities for developing new applications based on local inference.

Imran:

This is a huge benefit for privacy protection.

Qiao:

Not only that, but localized LLMs can also significantly improve response speed. Since all computations are done locally, there’s no need to send requests to the cloud and wait for a response. Even for people who don’t care much about privacy, the speed improvement is a major advantage.

Apple Intelligence & Privacy Issues

Imran:

This could be good news for Apple and Google, right? And for Android as well?

Currently, Apple is facing some challenges with its Apple Intelligence project. They had planned to launch some key features months ago and even heavily promoted them, but the launch was ultimately canceled. Reportedly, this was due to bottlenecks in privacy protection.

Apple can’t directly use models from other companies, so they have to rely on their own frameworks. But in this field, I think Apple is falling behind.

Qiao:

Apple is indeed lagging. What puzzles me, though, is that despite this, Apple’s P/E ratio remains over 30x, which is the highest among the “Big Seven Tech Giants,” second only to Tesla. Tesla, in essence, could be considered a meme stock. By the way, I bought some Tesla shares a few days ago.

But Apple’s valuation truly baffles me. They haven’t introduced any groundbreaking innovations in years. You could say Steve Jobs brought Apple a decade of brilliance, but after that, the shine seems to have faded.

Imran:

Let’s hope Tim Cook can turn things around. But Warren Buffett has already sold Apple shares at their peak.

Qiao:

With these high-performance, small-scale models becoming more prevalent—and most of them being open source—there’s immense potential for new applications based on local inference. This is hugely beneficial for both privacy protection and performance optimization.

Imran:

I think there’s massive potential here. We might see some very exciting applications emerge in the future. Do you know if any startups are exploring this space?

Qiao:

Not that I’m aware of at the moment. I think the biggest challenge right now is that when users download an app supporting local inference, they still need to download the model files separately, which could take a few minutes.

Using Localized LLMs

Imran:

I believe this issue will gradually be resolved as technology advances. This is definitely a space worth watching. I’ve already seen some potential use cases, such as applications in the health sector.

These models could help users analyze long-term health trends and provide personalized suggestions, like areas for improvement. And all this data would be processed entirely locally, without needing to upload it to a cloud-based LLM. That makes me feel more at ease. Sometimes, I don’t want to share certain sensitive information, especially for privacy reasons.

Qiao:

If there were an encrypted, fully private system where the data belonged solely to me, I would trust it much more.

Vibe Coding

Imran:

Have you ever used Lovable or other vibe coding platforms? I’m still trying to grasp the concept of vibe coding. Essentially, it’s an AI-assisted coding tool that can be applied across different domains. I recently built my own website, and I think these tools are truly magical.

Qiao:

These tools are highly competitive. Do you remember Wix and Squarespace? Those tools also helped people create websites quickly, but they were cumbersome to use.

Imran:

Exactly, the user experience wasn’t great. In contrast, Lovable’s standout feature is the freedom and customization it offers. For instance, you can easily tweak the design style to make the page simpler or remove images entirely. The operations are incredibly straightforward, and I feel these tools are now highly commercializable.

They’re suitable for almost everyone, especially entrepreneurs, because they significantly improve efficiency.

Qiao:

One of Lovable’s primary use cases is helping entrepreneurs quickly create product demos. For those who want to get a prototype into users’ hands without spending too much on development, this type of tool is incredibly practical. It’s also perfect for personal websites, like your own homepage.

Imran:

That’s fascinating because it lowers the barrier to entrepreneurship. As you mentioned on Twitter, we’ve noticed from data that today’s entrepreneurs tend to have lower technical skills compared to the past. This is because today’s “creators” can build applications without learning to code.

At this stage, the most important thing is to launch an MVP (minimum viable product) quickly and find a product-market fit.

Qiao:

True, but it’s worth noting that we still tend to favor teams with strong technical skills. That said, there might come a day when non-technical teams can quickly achieve product-market fit. I don’t know if that will happen by the end of this year, in two years, or five years, but it’s bound to happen eventually.

For now, though, we’re not quite there yet. What I’ve observed is that most teams today consist of two to three people, with one being a non-technical co-founder and the technical co-founder using coding-assist tools like Cursor and Windsor for front-end development. However, they still need a strong engineer to handle back-end development and tackle more complex tasks.

Imran:

Over time, products will become more complex. But at least for now, tools like Lovable or similar platforms can be relied upon to quickly build MVPs.

Qiao:

There have been some criticisms recently that vibe coding might lead to more errors in products. For example, some people are dissatisfied with tools like Cursor and Windsor. But for startups, these errors aren’t the critical issue. The biggest risk in the early stages is not having a product that users care about, rather than having a few bugs in the code. Once you launch a product that users truly need, you can always go back and fix those errors.

Imran:

Getting the product to market and into users’ hands is what matters most.

This can be especially impactful for entrepreneurs in emerging markets. For example, there’s a story about a kid from Congo who lived in a village without electricity. Using scrap metal, he built a wind turbine to generate power for the village. He learned and researched how to do it through the internet.

His actions can be seen as a form of real-world vibe coding—building energy solutions. From this perspective, emerging markets and local communities can use these tools to develop applications and solve real problems, improving their living conditions.

Qiao:

That’s because accessing funding is much harder in these regions, and their business models may not be suitable for venture capital.

Imran:

This excites me. I believe this will create a level playing field for more entrepreneurs, meaning we’ll see more innovative products that we’ve never encountered before. While I’m not particularly interested in the term “vibe coding,” the possibilities it brings are genuinely exciting.

Tokenizing Startups and Equity

Imran: How vibe coding relates to crypto?

There’s a startup called Taro Base—you’ve probably heard of it. The core feature of Taro Base is providing byte-coding tools to help crypto startups build applications, with a focus on crypto-related scenarios. We believe this field is entirely open for opportunities.

Beyond application development, Taro Base also enables people to tokenize startup ideas, transforming them into a kind of cyclical financing tool. This tool helps companies monetize while driving user growth. The concept of “turning your life philosophy into reality and eventually achieving an IPO” could become a new trend, and Taro Base is working toward this vision.

Additionally, there’s another startup exploring a similar model using games as an example. So, we’re already seeing some intriguing real-world use cases, especially in combining startup ideas with tokenization.

Qiao：

I really like the idea of tokenizing early-stage ideas and products. Essentially, this falls under the broader category of tokenizing startups, equity, or even entire companies, right?

So far, there seem to be two main approaches to tokenizing early-stage ideas on a broader scale: Directly tokenizing early concepts or tokenizing the stock of existing companies, such as putting Tesla or SpaceX shares on the blockchain.

The second idea is feasible but faces numerous practical challenges. First, you need to acquire stocks. Then the question arises: where do these stocks come from? You might need to purchase them as a company or acquire them from employees.

After that, you’d need to establish a legal and regulatory framework to put these stocks on-chain. This process is complex and riddled with friction.

Moreover, the market demand for this model doesn’t seem high at the moment because users can already access these stocks easily through their brokerage accounts.

So, another more promising approach is not to put secondary market stocks on-chain but to tokenize primary market stocks directly. In other words, companies would operate in a fully crypto-native way from their inception. For example, tokenizing their equity and putting it on-chain before the company even officially launches.

This could be the future direction for combining vibe coding with tokenization platforms.

Imran：

Essentially, this approach brings small and medium-sized startups directly onto the blockchain.

Qiao：

Exactly, and it also creates an entirely new market for these companies. I think that’s incredibly important.

BYD's New 1,000kw EV Charging Technology

Imran:

This could be a massive opportunity. Have you heard the latest news about BYD? As China’s largest electric vehicle (EV) manufacturer, BYD currently holds about 11% of the global EV market share, compared to Tesla’s 19%. Recently, BYD announced a groundbreaking technology that essentially allows an EV to complete a fast charge in just one minute.

Qiao:

How is that even possible?

Imran:

Their technology enables vehicles to achieve a full charge in just 5 minutes, providing a driving range of 250 to 300 miles. Specifically, they’ve developed a 1,000kW ultra-fast EV charging system.

The core innovation lies in combining platform flash-charging batteries with their signature blade battery design. This combination accelerates ion transfer within the electrolyte while reducing resistance in the battery separator, significantly shortening charging times. I don’t have a deep understanding of the technical details, but this breakthrough is precisely what has reduced the charging time to just 5 minutes.

Qiao:

Is this technology already in production, or is it still in the R&D stage?

Imran:

It’s expected to officially launch next month. What I love about this is how quickly they’ve moved from concept to real-world application. In contrast, I’m not as interested in conceptual technologies that take 18 to 24 months to materialize. This innovation comes from China. Another fascinating trend is that more and more cutting-edge technologies like this are emerging from China rather than the U.S.

Qiao:

I’ve seen some data indicating that Chinese car brands are rapidly gaining market share in Europe, especially over the past three years. However, I’m not sure if those numbers are entirely accurate.

Imran:

I recently watched a video comparing the interior designs of BYD and Tesla vehicles, and BYD’s interiors are far more refined.

I believe the competition in the EV sector isn’t just about the proliferation of charging stations but, more importantly, the capability of fast-charging technology.

Robotics

Imran：

I’ve recently been researching robotics, and it’s an incredibly competitive field. Robotics has immense potential across multiple industries. For instance, humanoid robots can be applied to various general-purpose scenarios.Factory robots specialize in improving efficiency, particularly in automated production lines.

Qiao：

This sounds like Amazon’s forte, right? While many people think Tesla is the leader in robotics, I believe Amazon is actually ahead. Amazon has been developing various types of robots for decades, especially in the factory robotics space, where they may already hold a significant lead.

Imran：

I agree with you. You’ve probably seen Andrew King’s post where he mentioned that robotics has a potential similar to the early days of cryptocurrency. He used Figure AI as an example to illustrate the massive opportunities in this field. I fully share his view—this is going to be a huge opportunity.

Interestingly, Elon Musk recently spoke about how the cost of goods may plummet in the future.

Qiao：

Yes, I remember he even predicted that the cost of goods could approach zero, which would reduce the need for money altogether. His logic is that when all goods are highly commoditized, currency would lose its purpose.

Of course, I think that’s a bit exaggerated—it feels more like a vision. But there are plenty of people who firmly believe in this idea.

Qiao：

Elon has a tendency to over exaggerate things. But even if it's just one tenth of that is a one 1/10 of 10 trillion, that would put Tesla at a valuation of maybe like 10 trilling the world's most valuable company.

Imran：

And they're gonna dominate generalized humanoids. Yeah, just based on what I've read so far. It's exciting.

To the workforce, both from an optimization perspective from factories, generalize use cases at homes. I'm starting to see surgical units that are leveraging robotics as a way to perform surgeries as an example, there's companies that are actually fundraised and have built these things.

Crypto and Robotics

Imran：

I’ve been researching the intersection of cryptocurrency and robotics, but so far, I haven’t found many practical use cases.

Payments seem like a possible application, but beyond that, I haven’t seen any clear examples of how cryptocurrencies could integrate with robotics. Do you have any ideas?

Qiao：

If there’s a potential use case, perhaps it’s in data-related applications. Like crypto incentives to bootstrap data? For training manual data, for training robotics.

Imran：

But do you think we really need cryptocurrency for that?

Qiao：

It's hard to say. Because the way Tesla did it with the with the Evs is they ship cars and then they capture the data with cars. And then it's like a self reinforcing cycle. They ship the actual hardware to capture data that is then used to improve the hardware.

Imran：

Tesla might adopt a similar strategy to advance humanoid robotics. My guess is that they’ll release an early version of humanoid robots to a select group of early adopters.

These robots might only be capable of completing 10–20% of tasks, with the remaining tasks requiring human assistance. Over time, robots could adapt to more general-purpose scenarios.

I believe this process would rely more on reinforcement learning than on crypto-based incentives. But who knows?

AI and Robotics in the Workforce

Qiao：

The thing I was thinking about the other day was all these like contractor freelancer platforms. Like Fiverr, Amazon, aren't they gonna die? Because of AI, because of the actual agents that do stuff. Before, we used to hire some contract freelancers to do some simple tasks like finding good founders on the internet, for example. Now you can just use those agents, do the same work in a much shorter time and cheaper.

Imran：

And in fact, I would argue that Fiverr is just a very difficult platform to use because most of the people that you're interacting with, obviously lower cost, is going to be from emerging countries. And because that there's a language barrier. So whatever you want, the output is actually the opposite. And you spend more time trying to convince the person to build what you want.