Original author: blockbeats
Original source: blockbeats
The 2024 Hong Kong Web3 Blockchain Week was held at the Hong Kong Convention and Exhibition Center from April 6 to 9. The event covered 4 venues and discussed the core issues of each track of Web3. In the past few days, the world's smartest and most innovative minds have gathered in Hong Kong to share and discuss the latest Web3 technical solutions. (For the specific event process, please read: "Event Express丨2024 Hong Kong Web3 Carnival will open on April 6, and the surrounding activities are summarized.") At the same time, many well-known investment institutions and crypto KOLs also participated. The event in Hong Kong has ended. BlockBeats editor has compiled several "small essays" of the Hong Kong Web3 Carnival for everyone to see their latest discoveries and insights, so that you will not miss the cutting-edge hot spots even if you don't go to the scene.
Crypto Weituo｜Crypto V (Partner of AC Capital):
I have returned to Myawaddy, and here is a belated essay about Hong Kong:
My observations:
1. This is a united and victorious conference. For the first time, I have the illusion that the industry has finally grown up. The VCs, ecosystems, project owners, and KOLs I met are more pragmatic than in previous years.
2. The theme of the conference is collusion and win-win, and there are "thought-out" conspiracy everywhere. Whoever co-organizes an event with whom on the stage represents who is in a group. This is no longer a shallow joint promotion, but mutual TVL, mutual liquidity, mutual top-up phone bills and free mobile phones.
Offstage, in the corridors of every event, in every private meeting, everyone finally stopped hiding for the so-called "brand tone" and spoke openly. I provided technology, you provided operations, he provided MM, how to divide the accounts and what market value strategy. The meeting finally began to have the efficiency of the Canton Fair.
3. Meet up with old friends and KOLs to discuss the advancement and promotion of agricultural technology. The hypothesis of "KOL replacing VC" proposed two years ago has finally been realized in a practical sense. The title of KOL can be used openly, and more so-called "international major projects" have begun to adopt the direct sales financing strategy of "two listed VCs + exchanges + KOLs".
4. The conference has 5 main themes: using BTC as a bet chip, using Solana as a bet chip, mutual aid Restaking looking forward to the split stage, the new Cosmos public chain-level Ponzi scheme, and TON’s promotion on Tencent Conference.
my thoughts:
1. Crypto’s opportunity window is not narrowing, but getting bigger
- There is no consensus on the main line between the East and the West anymore, and the liquidity war will reach a new level. The "political correctness" of the last round no longer exists, and each camp has put out a large budget to find new heroes.
- The consensus and participation in on-chain opportunities have increased significantly. Shanghai Stock Exchange is the ultimate goal of collusion, but it is not the only solution.
- The theory that "grassroots people have no chance" has been proven wrong. What is eliminated are "grassroots professionals" rather than "professional grassroots people". The experienced local dog team is more popular with the conspirators than the out-of-touch "title stamp collectors".
- "Compliance" is just superficial political correctness. Everyone's products and efforts have already explained the problem. As the global political situation is divided, regulations continue to increase, but compliance will become increasingly weak. I admire the courage of those teams that have gone from Silicon Valley to Dubai and cut ties with the powers of the industrial age.
2. All VCs work for Binance
Those who can't work will inevitably lose their position. Accordingly, either do what Binance likes, or do something completely opposite. Going to extremes will become an advantage in this cycle. As @BroLeonAus said, the lack of exit paths for institutions will bring opportunities to @Bybit_Official@gate_io@bitgetglobal. In fact, I am more looking forward to the emergence of new vertical centralized institutions in this cycle, especially newcomers in Southeast Asia or Russia, which are sanctioned or have no camp. Everyone stopped pretending to be halal in this cycle, can they stop pretending to be compliant in the next cycle?
3. Reflect on yourself
During this trip, I had in-depth exchanges with many colleagues, and I realized that I had a visible gap in the professionalism and coordination of team cluster operations. Old sickle, new problems, it's time to make up for the shortcomings. New cooperation, new conspiracy, new organization.
4. Open Source Scythe
I met many friends who like open source sickle, which made me realize that the "three-disk theory" is appreciated and even really helps many people. This is the right way, and I will continue to practice, summarize and share.
Let there be no difficult plate to open in the world.
grateful
5. Keep fighting
The bull market is far from over, or is it no longer appropriate to view the industry we are in in terms of bulls and bears? As an investor or a market maker, should we wait, rely on, and demand with the "Bull and Bear divide" in mind, or should we say that this market should be a "fight to win or lose" one after another for us?
Original link
Haotian (Cryptography Researcher):
After a three-day trip to the Hong Kong Web3 Conference, I have a lot of feelings, but I can’t express them in words. I will just use the spare time to share my feelings:
1) From the excitement of the conference, it is certain that the market is in the early stage of a bull market. We can feel the strong enthusiasm for participation, but it is not too Fomo yet. Although the secondary market has been volatile during this period, the expectation of the Fed's interest rate cut has been delayed, and the uncertainty of the bull market has increased, the enthusiasm of the primary market builders to participate is certain, which is also the basis for this round of bull market to be firm, lasting and magnificent enough;
2) The theme sharing and panels at the main venue and sub-venues, as well as the surrounding community-led activities, etc., are decentralized, thematic, and community-based, so it is difficult to say which one is the mainstream. Some people believe that Solana can flip Ethereum, some believe that CKB can drive the prosperity of BTC Layer2, and some people go to the AI+DePin direction to find Alpha. Of course, more people stick to the Ethereum Layer2 field. It seems that everyone has their own "mainstream" in their minds, but there is no one that has gained national consensus and the "main uptrend" of national Fomo has appeared. The entire northwest of Shanxi has become a mess, and who is the leader has not yet been decided. This just proves that it is still in the early stages of the bull market, and there is no track or leader with absolute voice.
3) After socializing outside the venue, I felt that the weight of CEX exchanges was not so "prominent". Maybe they were forced to hide in the dark for some reason. But without the rows of beautiful Western models and the wide range of Shill visual effects, I could absorb more nutritious spiritual food. Although there are still many kinds of technical, community and industry narratives that dazzle and confuse people, it is obvious that the stories are being told seriously and the market aesthetics are also raised. After the Bitcoin ETF was passed, it is not only BTC itself that has been mainstreamed, but also the Build mentality of Crypto practitioners? In fact, the market is not afraid of telling stories, but is afraid that the market treats leeks as fools.
4) The narrative around "chain" is still the top priority, whether it is ETH Layer2, BTC Layer2, high-performance Layer1, parallel EVM, modular chain, chain abstract interoperability, AI+Crypto, ZK, DePIN, etc., each narrative direction seems to be centered around "chain" infra. On the surface, all parties in the industry have focused their development efforts on the narrative of "chain", and the application landing direction seems to have stagnated? But in fact, the application landing direction is also very lively, but the narrative of the application can only have a sense of existence if it reaches the same phenomenal level as Stepn. Once the application is locked, it has to go through Internet-style involution such as attracting new users, retention, and growth, and the current market's Mass Adoption level is far from the point where applications can be rolled. On the contrary, rolling B-side services can be nested with modular resources and combined with liquidity across the ecosystem with chain abstraction. The scalability is really rich. Rather than saying that the market prefers to tell B-side stories, it is better to say that this is all we can tell at this stage;
5) It can be clearly felt that the new love of BTC Layer2 and the old love of Ethereum Layer2 are engaged in a competition. The Layer2-centric story vision proposed by Vitalik is very beautiful, but it is still unsatisfactory when reflected in the currency price and actual delivery expectations. It is precisely because of this , BTC Layer2 has only taken on large enough new market expectations. Therefore, there is no doubt that BTC Layer 2 is reaping the dividends of Ethereum Layer 2 being less than expected. Of course, this kind of involution will be suspected of reinventing the wheel, but don’t ignore it. The user groups who play in the BTC ecosystem and those who play in the Ethereum ecosystem are not 100% overlapping. The strong consensus of BTC is likely to be the core driving force of this bull market. . In the long run, BTC Layer2 can derive a unique new route of UTXO architecture, and can also be combined with the mature Ethereum layer2 ecosystem. If we take Layer2 as the basis, BTC Layer2 and Ethereum Layer2 are not a life-and-death competition. It can be clearly seen that It feels like both camps are embracing each other;
6) We have indeed not seen any excellent Alpha projects that are very innovative and inspiring. Some people say that this is a stage for Eastern forces to show off their muscles, and Western forces are all concentrated on Dubai Token2049 a few days later. In terms of narrative direction, AI, DePIN, ZK, and parallel EVM, if you don’t consider landing, Western forces are indeed more attractive. After all, the valuations of many Western projects are terribly high when they are spread to the Chinese-speaking circle; in terms of innovative vitality, the inscription market, BTC Layer2, and DePIN hardware mining, if you don’t care about Rug friction, the wealth effect of Eastern forces in attracting new users is indeed driving the entire industry forward. Rather than saying that Eastern and Western forces are competing secretly, it is better to say that each has missed each other’s growth. The current Eastern inscriptions and Western MEME forces will always form a combined force after a certain main rising wave force is generated.
7) It is obvious that the voice of high-quality KOLs has become stronger, and to some extent, their influence has surpassed that of some VCs. This is fundamentally due to the change in the way VCs play. VCs work with project parties to incubate projects and attract exchanges to accumulate projects. This round of VCs has obviously become "old leeks". The VC dream-building story that has locked up for a long time in the past no longer works. To some extent, the money that wants to escape at any time has become a hot potato. Most excellent projects cannot get quotas just by having money, and KOLs who have both primary investment capabilities and secondary sales capabilities have become popular. Therefore, the KOL Round did not come out of thin air, and it is a good supplement to the weakness of VCs' weakening voice in the primary market.
8) Under the guidance of narrative directions such as Restaking and modularization, the resource integration (La Bang Jie Pai) capabilities of the project have become increasingly prominent, which is tantamount to the manifestation of soft power beyond the mainstream technology and development narrative of the project. It should be said that the resource accumulation ability behind the project will become a strong support for its development expectations. Once a project has extremely strong resources and relationships, judging from a business perspective, it has already won at the starting line. When telling the same story that is difficult to implement, which type of project will users prefer? The answer is obvious;
9) There are too many stories of old trees sprouting new buds in this round, and I personally like to see them. Although new chains also have the conditions and environment for one-click launch, the Great Again of old chains hides the earnest expectations of each community and the expectations that should have been delivered. If old chains can tell new stories, revitalize the community and consensus, and keep pace with new chains, I personally hope that more old chains will rise.
10) There will be all kinds of FUD voices in every meeting, just treat it as gossip. TVL false prosperity, DeFi mining income dilemma, potential exchange crash, etc. If you agree that resource integration is the current market theme, FUD TVL fraud, which is tearing off the industry's bottom line, has no meaning. As for black swan events under uncertainty, it is difficult to avoid them, but generally speaking, projects and teams that are capable of crossing bull and bear markets will be able to move more steadily and maturely in this round. Fud cannot kill a project, it can only increase some market noise.
Note: The above thoughts and ideas only represent my personal experience and observations at the conference, and cannot reflect the full picture of the Hong Kong Web3 Conference. However, I hope that the industry can tear off the labels of niche and scam, and gradually develop in the direction of being popular and mature. I really hope that one day I will not understand this industry, but I can always feel its growth.
Original link
Fred the Scallion (crypto investor):
After coming to Hong Kong recently, I talked with some investors and found that everyone is paying close attention to the AI and DePIN tracks. I would like to share my recent thoughts on AI and DePIN and discuss with you.
Four core issues were discussed:
1. Why do most decentralized computing power projects choose to do AI reasoning rather than AI training?
2. What is Nvidia's strength? What makes decentralized computing power training difficult?
3. What will be the final outcome of decentralized computing power (Render, Akash, etc.)?
4. What will be the final outcome of the decentralized algorithm (Bittensor)?
Next, let's unravel the layers:
1) Looking at this track, except for Gensyn, most decentralized computing power projects choose to do AI reasoning rather than training. The core reason lies in the different requirements for computing power and bandwidth.
To help everyone understand it more simply, let's compare AI to a student:
- AI training: If we compare AI to a student, then training is similar to providing AI with a large amount of knowledge and examples, which can also be understood as data, and AI learns from these knowledge examples. Since the nature of learning requires understanding and memorizing a large amount of information, this process requires a lot of computing power and time.
- AI reasoning: What is reasoning? It can be understood as using the knowledge learned to solve problems or take exams. In the reasoning stage, artificial intelligence uses the knowledge learned to solve problems, rather than activating new knowledge, so the amount of calculation required in the reasoning process is relatively small.
It is easy to find that the difference in difficulty between the two is essentially that large-model AI training requires a huge amount of data and a very high bandwidth requirement for high-speed data communication, so it is currently very difficult to implement decentralized computing power for training. Inference, on the other hand, requires much less data and bandwidth, and is more likely to be implemented.
2) So where are the bottlenecks for data and bandwidth? Why is decentralized training so difficult to achieve?
This involves two key factors in large model training: single-card computing power and multiple cards in parallel.
- Single card computing power: At present, all centers that need to train large models are called supercomputing centers. To facilitate everyone's understanding, we can use the human body as an analogy. The supercomputing center is the tissue of the human body, and the underlying unit GPU is the cell. If the computing power of a single cell (GPU) is very strong, then the overall computing power (single cell × number) may also be very strong.
- Parallel connection of multiple cards: The training of a large model can easily be hundreds of billions of GB. For a supercomputing center that trains large models, at least 10,000 A100s are needed. Therefore, these tens of thousands of cards need to be mobilized for training. However, the training of large models is not a simple series connection. It is not that the training is completed on the first A100 card and then on the second card. Instead, different parts of the model are trained on different graphics cards. When training A, the results of B may be needed, so it involves parallel connection of multiple cards.
Why is Nvidia so powerful and its market value has soared, while AMD and domestic companies such as Huawei and Horizon Robotics are currently struggling to catch up? The core is not the computing power of a single card itself, but two aspects: the CUDA software environment and NVLink multi-card communication.
- On the one hand, it is very important to have a software ecosystem that can adapt to the hardware, such as NVIDIA's CUDA system. However, building a new system is difficult, just like building a new language, and the replacement cost is very high.
- On the other hand, there is multi-card communication. In essence, the transmission between multiple cards is the input and output of information. How to connect in parallel, how to transmit. Because of the existence of NVLink, there is no way to connect NVIDIA and AMD cards; in addition, NVLink will limit the physical distance between graphics cards, requiring the graphics cards to be in the same supercomputing center, which makes it difficult to achieve decentralized computing power if it is distributed around the world.
The first point explains why it is currently difficult for AMD and domestic companies such as Huawei and Horizon Robotics to catch up; the second point explains why decentralized training is difficult to achieve.
3) What will be the final outcome of decentralized computing power?
- It is currently difficult to train large models with decentralized computing power. The core is that the most important thing for large model training is stability. If the training is interrupted, retraining is required, and the sunk cost is very high. It has very high requirements for multi-card parallel connection, and the bandwidth is limited by physical distance. NVIDIA uses NVLink to achieve multi-card communication. However, in a supercomputing center, NVLink will limit the physical distance between graphics cards, so the decentralized computing power cannot form a computing power cluster for large model training.
- But on the other hand, it is possible to meet the requirements of relatively low computing power, such as AI reasoning, or small and medium-sized model training in some specific scenarios. When there are relatively large node service providers in the decentralized computing network, there is potential to serve these relatively large computing power requirements. And edge computing scenarios such as rendering are also relatively easy to achieve.
4) What will be the final outcome of the decentralized algorithm model?
The final outcome of the decentralized algorithm model depends on the final outcome of the future of AI. I think the future AI war may be that there will be 1-2 closed-source model giants (such as ChatGPT), plus a variety of models. In this context, application layer products do not need to be bound to one large model, but rather cooperate with multiple large models. In this context, Bittensor's model has great potential.
Original link
Dov Wo (crypto investor):
A short essay on attending a conference in Hong Kong from a VC perspective (full of bias):
1. In a bull market, products are not important, narratives and emotions are more important. Meme has repeatedly proved this. At least half of the money raised is used to boost the market.
2. The most core abilities in this bull market: trading ability and shouting ability. If you don’t have trading ability, please practice shouting ability. I am also helping some project parties find KOLs for promotion. Welcome to take my orders. There are many projects that Binance and OKX have invested in but cannot find suitable domestic/overseas promotion channels. False prosperity/buying is just a necessity.
3. The status of VC is getting worse and worse. The unlocking is slow, the valuation is high, and retail investors do not recognize it. If they cannot bring additional resources to the project party (such as exchange relations, resources and communities in a specific region, economic model design, promotion capabilities, etc.), most pure financial investment VCs can only be a big complaint.
Take a recent project that approached me as an example. The valuation for the KOL round was lower than that of VC, and the unlocking was better than that of VC. (This project was led by a top Western VC)
KOL > VC is not groundless, it really exists. So I am also running a small boutique KOL agency to sell shovels.
4. The most popular activities are Berachain, Solana, and BTC; the ETH ecosystem is relatively inactive, and a big meme is Layer 69 (Solana’s own funny video making fun of ETH)
5. There were a lot of people at the Side Event. There were probably hundreds of Side Events at this event. On the contrary, there were not many people at the main venue. Project owners may need to reflect on whether the expensive booths at the main venue are still worth it. For example, Berachain and Solana did not have booths at the main venue, but held their own events. OKX had booths and events, which was the top status. On the other hand, Binance still had no events.
6. Ordinary people must become KOLs and find their own position. For example, if you want to shout "Tugou", just shout "Tugou". If you do a good job of analysis, just do a good job of analysis. You can get good business orders. I asked KOLs who shouted "Tugou" to shout orders, and the price of each order ranged from 200U to 2000U. VCs who can't shout orders can't survive, and KOLs who can't shout orders can't make money. If you have traffic, a large number of project parties will actively look for you to cooperate, and you will have better information and more resources. Please read the second point again. Question the teacher who leads orders, understand the teacher who leads orders, and become a teacher who leads orders.
7. This year, HK talked less about compliance and licenses, and more about trading and accumulating resources (maybe because I have reached a higher level and can understand and participate). The largest booths as soon as you enter the main venue are OKX and DWF. Everyone knows who has resources and who has money.
ABCDE founder Du Jun commented on the tweet, saying that "it will be difficult for institutions without core investment and research capabilities."
Original link
Ethan Yu（AC Capital Partner）：
After drinking it for 2 days, I listened to some awesome stuff:
1. Chinese institutions Fomo Bitcoin ecology and Restaking, European and American/overseas institutions Fomo DePin on Solana, a small number of institutions Fomo Cosmos ecology-related Bearchain, some institutions were trapped in GameFi in the last round, so now they can only continue to push forward, but all of them end up in the same place with big cuts, and the valuations are simply outrageous.
2. Almost all consensus is that there are too many Golden Dogs on Solana, and the get-rich-quick effect has been continuing. In the past few days, the flow of local Dogs that went to Base has returned to Solana.
3. It’s not that KOLs play a bigger role than VCs, it’s because KOLs have more flexible funds and can call orders and delete pushes at will, and the unlocking method is also more friendly. What can you do if institutions call orders? On the other hand, institutions are not stupid and have a lot of money. You start with a valuation of 100 million US dollars, and it takes 3 years and 36 months to fully exit. You don’t list on Binance or OKX. If you let me invest 500,000 US dollars now, are you stupid or the institution stupid?
4. There are indeed many newcomers and people who want to enter the industry. Young people are proactive, enthusiastic, and eager to learn. They have no industry baggage and are willing to do it. However, there are not enough newcomers, so the bull market is still in its early stages.