Facta non verba.

LFG Backpack 🎒one of my most fav portco

https://x.com/danpark012/status/1876581007518024080?s=46

BM

Boyco 1B

https://dune.com/thj/boyco-markets

Announcing Hashed Thesis 2025: Leading Blockchain’s Global Adoption with Asia at Its Core

Full Medium article here

We are excited to announce our co-lead investment in GAIB, a project revolutionizing the compute asset ecosystem through tokenization and financialization. — Link

2nd panel after AiFi Summit panel (pretty chill wrt panel schedule for me this time in BKK 😁)

Speaking at DePIN Day at Devcon 7. Hmu if you’re around! — Luma link

Sequoia Capital and the evolution of the VC industry | FT Film — YouTube video from FT

- Founded in 1972 by Don Valentine, Sequoia Capital is a premier VC firm in Silicon Valley with a 50-yr history.
- It has invested in major companies like Apple, Google, Oracle, Cisco, PayPal, Airbnb, YouTube, Instagram, and OpenAI.
- Notably, companies backed by Sequoia make up about 25% of the Nasdaq index, highlighting its significant impact on technology and innovation.

- In 2022, Sequoia invested $225M in FTX. This raised concerns about DD practices in VC investments.
- The loss was especially notable as Sequoia had been more cautious about crypto compared to rivals like a16z, which had aggressively invested in the sector.

- Sequoia split its global operations into separate entities for the US & Europe (remaining as Sequoia), China (now HongShan), and India (now Peak XV).
- The Chinese branch was managing over $50B in capital at the time of the split.
- Sequoia still holds stakes in Chinese companies like ByteDance, the owner of TikTok, which is facing potential bans in the US, potentially affecting shareholder value.

- Generational leadership changes led to internal disputes within the firm.
- Notably, at Klarna—a BNPL company that was once Europe's most highly valued startup—Sequoia partner Matthew Miller attempted to remove former managing partner Michael Moritz from the board.
- This public conflict highlighted tensions between current management and former partners during the leadership transition.

- The VC industry has expanded, with the number of firms in the U.S. quadrupling over the last 10–15 years, leading to increased competition and inflated valuations.
- Sequoia is adapting by heavily investing in AI, allocating 60% of its current investments to the sector.

Just refreshed the disclaimer as I go live again

A deeper look into the potential of NodeFi & GPU financing — Article from Niklas Polk at Nansen

- The booming AI sector offers lucrative investment opportunities through GPU rentals via DePIN, providing significant USD yields.
- Concurrently, NodeFi delivers high non-USD yields.
- Cryptocurrency and blockchain enhance the GPU rental value chain by enabling financing, fundraising, and unique yield distribution through token emissions and DeFi integration.
- Potential annual yields can reach ~200% in optimistic, ~95% in realistic, and ~30% in pessimistic scenarios by combining asset returns, token emissions, and DeFi yields.
- Emerging DeFi products like GAIB and Aethir are essential for developing on-chain GPU financing.
- This approach mirrors the rapid growth of Bitcoin mining, with innovative financing models expected to accelerate the crypto-adjacent GPU sector.

Past few months meme rotation & Some learnings — X Thread by @ahboyash

- Stay updated on meme trends, like recent favorites such as Base memes, AI agents, and Gen Z phrases.
- Build a consistent investment strategy and be selective, focusing on $5–10M caps for better returns over low caps.
- Investments should be thesis-driven, based on cultural or influencer appeal, and capital should be readily available on multichai.
- Tracking Dexscreener Multichart helps monitor trends by ecosystem.
- With the market evolving, old strategies are less effective, and adapting to the current meta is crucial.
- Young investors should beware of “brain rot” from chasing meme trends without developing long-term skills.

My conclusion - Last comment so crucial... If I'm not confident I can monitor my meme position every single day, will not try to ape in with my own capital... unless it becomes mainstream hard to ignore

Part 3: Is NVIDIA's Era Coming to an End? Formation of the Anti-NVIDIA Alliance and Glass Substrates (UXL, UALINK, Glass Substrate) — Link

- NVIDIA, while leading in GPU-based computing power boosted by hyperscaler investments, faces new market variables due to design flaws causing delays in its Blackwell series. Growing dissatisfaction over NVIDIA's high prices and closed ecosystem (CUDA, NVLink) is prompting companies, including hyperscalers, to seek alternatives.

- The UXL alliance, based on Intel's OneAPI, is building an open ecosystem alternative to CUDA, addressing past shortcomings like hardware manufacturer support. UALink is developing an interconnect standard to rival NVLink, aiming to reduce NVIDIA's lock-in effect by allowing compatibility among different suppliers' hardware.

- The rise of on-device AI is expanding the edge device-based AI computing market, forming new areas beyond NVIDIA's influence. There's a shift in performance enhancement focus from design to manufacturing processes, advanced packaging, and materials like glass substrates, potentially weakening NVIDIA's core competitiveness.

- Glass substrates are overcoming limitations of existing organic substrates, offering higher performance and efficiency. Korean companies like SKC, Samsung Electro-Mechanics, and LG Innotek are leading the market through active investment and research in this field.

Part 2: Google, Apple, and Microsoft in Crisis; The Birth of SearchGPT; What Is OpenAI Aiming For? — Link

- There's an evolution away from exclusive cloud-centric systems toward additional embedding algorithms within edge device, which enhances personal data protection, security, and resource efficiency. Federated learning allows AI models to be trained on individual devices, sharing only updated parameters. Tesla's 'Dojo' is a prime example, where vehicles collect real-time data to continuously update AI models.

- Companies like Google, Microsoft, Apple, and Amazon are adjusting their strategies in line with this shift to distributed models. Despite possessing massive data, they face privacy constraints; federated learning offers a solution to enhance AI competitiveness while addressing privacy concerns.

- AI supplements real data by generating synthetic data, filling learning gaps especially in areas where data collection is difficult. The competition for AI dominance hinges on effectively handling federated learning and synthetic data. Companies leading in these fields, such as OpenAI with its recent "SearchGPT" release, are poised to shape the future AI ecosystem.

- Ultimately, future dominance in AI is likely to be determined by data access methods and learning efficiency.

Although it's a two-month-old video, I'm sharing a three-part series that clearly summarizes the current state of algorithms, big data, and compute in AI. While some points might be familiar, the message is clear, and I recommend watching when you have time (total runtime: about 40 minutes).

All videos are in Korean btw...
Part 1: Algorithm (Link)
Part 2: Big data (Link)
Part 3: Compute (Link)

Part 1: Google, Apple, and Microsoft in Crisis! The Calculations of Top-Tier Companies in an Era of Great Transformation — Link

- The AI era is centered on algorithms, big data, and compute, marking a transition in the fourth industrial revolution from 'Tool' to 'Machine', 'Device', and finally, 'Function'. LLMs like GPT are breaking down language barriers and enhancing human-machine interaction.

- With the advent of on-device AI, algorithms are now executed directly on devices rather than relying solely on the cloud. This enables the dynamic selection and processing of algorithms based on user needs, providing personalized AI experiences.

- In the AI era, companies that control the selection of algorithms, rather than just their creation, are poised to seize market dominance. Platform companies like Apple, Google, and Microsoft hold significant positions due to their ecosystems. For example, Apple leverages years of user data and queries within its ecosystem to provide AI services optimized for user needs; the development of on-device AI further enhances this capability.

- The intersection of LLMs and GenAI, along with on-device AI that selectively executes these models, is creating new markets led by major platform companies. These companies are likely to steer market trends through their authority to select algorithms.

UBS Asset Management launches its first tokenized investment fund — Official announcement link

- UBS Asset Management launched the “UBS USD Money Market Investment Fund Token” (uMINT) on Ethereum, offering tokenized cash management through authorized distributors.
- Growing demand for tokenized assets has led UBS to expand blockchain-based fund solutions, aligning with MAS’s Project Guardian.
- In 2023, UBS issued CNH 200 million in digital structured notes and completed a cross-border repurchase transaction with a digital bond on a public blockchain.

Current status of VC's infrastructure funding — X link

Introducing Based Agent: create AI agents with full onchain functionality on Base in less than 3 minutes. — X announcement link

- Replit: https://replit.com/@KevinLeffew1/Based-Agent?v=1
- CDP SDK Docs: https://docs.cdp.coinbase.com/mpc-wallet/docs/welcome
- GitHub repo: https://github.com/murrlincoln/Based-Agent/tree/main/Based-Agent