Poseidon @psdnai 宣布完成 1500 萬美元融資，由 a16z crypto 領投，背後是 @StoryProtocol 孵化，這組合不小，看起來是真的要往 AI 數據基建方向搞了。 簡單說明一下兩者的分工： •Story Protocol 主要解決「IP 授權與確權」的問題 •Poseidon 則聚焦在「實體世界的數據採集」 一個負責數據的產權歸屬，一個負責數據的生產流通，有點像是鏈上版的 Nvidia + ScaleAI。 現在 AI 模型餓得不是算力，而是「真實世界的訓練數據」。 什麼車上的鏡頭、倉庫裡的監控、語音對話、掃地機器人的路徑紀錄⋯⋯這些才是 AI 模型真正要吃的料，而這些數據，Poseidon 想把它們包裝成可交易的資產。 他們的模式也不複雜： 1.幫用戶收集並打包日常數據 2.幫企業做資料標註與處理 3.把使用授權放到鏈上 4.然後提供分潤給原始數據提供者 簡單來說，就是讓「生活的數據」也能變現，AI 公司為了訓練模型，得乖乖付錢來買。 這種項目未必像 meme 幣一樣爆衝，但有可能成為 AI 時代底層的「數據煉金基建」。 說白了，炒幣太老派了。 接下來，可能要開始學怎麼「炒數據」了。

Post-mortem on $PUMP: - Was very bullish at 4bn and moderately bullish at 6.5bn. Was looking to size X% at 4 and about 0.6X at 6.5--this was mostly fine IMHO given the extent of demand during the ICO. - Existence of pre-market perps changed the dynamics quite a bit. It became hard to differentiate "true" ICO demand vs. hedging demand--how much of the buying/oversubscription was looking to hedge vs outright long. I overindexed on the demand and thought the pre-market price was actually somewhat suppressed given hedging flows were pushing it down. This was very wrong--sustained positive funding on HL pre-launch and all perps post-launch indicated otherwise. - Believed most of the spot longs that weren't long-term holders would've hedged out around 6-7bn given favourable funding + extent of time it traded up there + closeness to most people's targets. This was wrong--tough to know a priori. - As a corollary, this trade became incredibly crowded. Every participant was long with the same target (8bn ez). Crowded/consensus trades are fine to take as long as further flows are coming (beneficial if those additional flows are exogenous to the cohort that is crowding). Smart-ish and regular money had already allocated to PUMP via perps even after ICO and there was no chance outside dollars were going to come into this product. @AlwaysBeThinkin about your exit liquidity. - My price targets were looking to derisk between 7-8 given the memetic consensus around 8. This was too aggressive and that 7-7.2bn rough resistance level from pre-market was a good level to take some off. - Risk-on markets are good for beta, but SOL was locally lagging and ETH chadding double digit percentages per day sucks attention and liquidity further away from a new launch that had to work through substantial ICO supply. I am 90+% out of the position now. Think it has legs to run but opportunity cost of holding this is too high in this environment. I averaged in around low 6s and closed most around mid-4s. Took a big drawdown on this; was able to cover it and more given the current environment but put a sizable dent in the book. h/t @0x_Leo_ - he wrote some very good reflections that inspired me to share mine. Onwards and upwards.

Postdam's #EISAPEC23#, here we are! Looking forward to 2 presentations in panels + 1 "authors meets critics" + 1 discussant role. All on #populism# in/and International Relations! Write me if you want to chat about these & other issues! @europeanisa https://t.co/VEryvFNtpH