Carl Vogel

Carl Vogel@CarlKVogel · 2 天前观点输出

RT @0xSherpaXYZ: We love Polymarket and settle 100% of trades there, but their taker fees have gotten pretty steep. We route ~ 90% of our…

AI：延续个人公开立场，强调长期主义、执行效率与行业方向判断。

Carl Vogel@CarlKVogel · 2 天前赛道影响

And this is why public blockchains will struggle to get institutional adoption in their current form (plus MEV)

AI：偏主题投资表达，释放对基础设施与新叙事的偏多信号。

Carl Vogel@CarlKVogel · 3 天前政策影响

This was a fun CryptoxAI Panel! Sharing some top-level raw notes from the panel -------- The multi-trillion dollar question is, 'If/How will Crypto make AI more useful and impactful?' For all the hype of Crypto x AI, today there's not much overlap between them. In-fact the two largest CryptoxAI companies have pivoted to be just ... AI. We've reached peak AI centralization. Enterprises realizing that a multi-model approach which leverages a combination of frontier-lab and open source models is more capital efficient and better protects proprietary data. There's therefore a push to 'on-prem'. But with the growth rates of AI models and the therefore exponential need for more GPUs, is an on-prem strategy realistic for most companies? Probably not. Neo-clouds will win. But I also think there's a reemergence opportunity for blockchain-based inference networks, they can both secure data / provide anonymity while also providing better global coverage and dynamic loads. The key challenge will be, can a decentralized network connect enough GPUs to run the higher-parameter models? Future areas for real intersection Agentic Payments Agents are currently 'single-player' mode; each of us may have an agent, but they don't really interact with each other. Until they do, will be limited adoption for agent-to-agent payments and protocols. When we do get to that point, open-protocols like blockchains will win over closed networks because they are programmable, immediate, and global. Today agentic commerce is giving a debit-card to your Agent w/ spend limits Agentic Trading Agentic trading will be huge. Today trading UIs today are designed for HUMANS, which are lazy, impatient and emotional. Agents enable completely different trading experiences. Agents can better optimize strategies than humans, doing consistent rebalancing and risk-mgmt for a given objective. Even for simple use cases such as 'buy sol', an agent can tax-loss harvest that position and rotate it into a covered-call strategy in a sideways market, generating additional income on the underlying position. Humans cannot be presented with too many options, but Agents can ingest infinitely more data 24/7 than humans. One question was, does agentic trading make finance more reflexive, higher vol? Some believe yes, that most AIs (especially if not well prompted) will look for similar heuristics, meaning more plowing into trades at similar times and reading and trading on the same signals. But my initial counterpoint is that Agents are smarter than humans, so less copy-trading and 'impulsive' or 'emotional' trading than what we see today in markets, which is VERY reflexive. Market Innovation (eg Compute + Data Markets) Crypto enables any market to be created and distributed globally instantly, much more quickly than tradFi, regulated approaches. For example, H100 compute markets (pricing) have been trading on Hyperliquid for months, and only now approaching going mainstream through Kalshi and others. Also, market creators don't need to build an orderbook (eg CLOB), simply need to build the oracle or index, much lower lift and faster time-to-market. Many more markets because they don't need to support the regulatory and tradfi overhead Data Provenance / Trusted Data Huge overhead for institutions/enterprises to verify AI outputs since they don't know the data the model was trained on. Much cheaper to log data provenance on the way-in than verifying all AI outputs, or at minimum can source trusted data in the inference response (even if not trained on that data). Big future opportunity for enterprises to train their own smaller models w/ their own trusted data, reducing latency on outputs and required context windows

AI：偏政策推进，强调合规落地与规则明晰，对监管主线更敏感。

Carl Vogel@CarlKVogel · 4 天前观点输出

RT @jtx_trade: JTX is now live for everyone. Trade any asset on Solana, from memes, tokenized equities, majors, and more, with a trading…

AI：延续个人公开立场，强调长期主义、执行效率与行业方向判断。

Carl Vogel@CarlKVogel · 4 天前观点输出

RT @Pumpfun: Introducing BOOST mode - the new standard launch mechanism for EVERY new pump fun coin Over $100M in dead liquidity is lost e…

AI：延续个人公开立场，强调长期主义、执行效率与行业方向判断。

Carl Vogel@CarlKVogel · 4 天前观点输出

Always about Apps

AI：延续个人公开立场，强调长期主义、执行效率与行业方向判断。

Carl Vogel@CarlKVogel · 4 天前观点输出

Any traders (sports, prediction markets, equities) looking for deeper abilities to create ML models? Cool company that's enabling anyone to easily spin-up ML models and looking for intros to early users

AI：延续个人公开立场，强调长期主义、执行效率与行业方向判断。

Carl Vogel@CarlKVogel · 7 天前观点输出

@marccolcer Who cares!

AI：延续个人公开立场，强调长期主义、执行效率与行业方向判断。

Carl Vogel@CarlKVogel · 8 天前观点输出

"Nothing good happens below the 200-day moving average" The clearest indication that "we're back" is when $BTC breaks through the 200d SMA. Same for $SOL. You can bid now to catch the wave early for a better return, but the largest institutional capital is watching how we re-interact with the 200d. You can see in May we were definitely rejected and trended lower. Currently those levels are roughly $72k for BTC and $88 for SOL

AI：延续个人公开立场，强调长期主义、执行效率与行业方向判断。

Carl Vogel@CarlKVogel · 8 天前观点输出

@therealchaseeb @vibhu There are levels to this 🫡

AI：延续个人公开立场，强调长期主义、执行效率与行业方向判断。