Jason Kam

Anthropic claims Claude Opus 4.6 published today is a marked improvement for investment analysis; I put it to the test. I ran it through my automated stock pitch pipeline: 6 AI analysts, 5 rounds of debate, 1 research director. It cost ~$50 (20x my usual). Is it worth it? Here's what I found: 🧵

@silverfang88 还有雄安，看起来像下注新领导的投名状。

2026年拼多多花了33亿 买了陆家嘴星展银行大厦 2006年搜狐花了22亿 买了五道口和中关村的两栋楼

Hmmm what does that mean for $SPCX bruv

国内可回收火箭成功了，卧槽，牛逼炸了。 @elonmusk

~70-80% of GEOD tokens are sold (assume moved = sold) within 6 months. Probably safe to model the same going forward as well.

Fwiw $GEOD had been compounding its ARR at 100%+ annualized rate and now sits at 11 m, or about 750k / month dedicated to buyback. Post halvening the 20k+ stations get 6 GEOD per day, or 700k USD / month at current price. The protocol is now at net burn. Assuming $GEOD ARR hits 30 mm as per Kyle’s indication in a year, after another halvening with price staying the same, the protocol would be burning ~15 mm USD per annum, or ~14x forward burn. As long as team executes and ARR keeps ramping, I expect $GEOD to continuously grind up from here on out.

Man this Fable-5 is a fucking horse

@hash_dd It’s 0.25 per my math currently after this halving.

@MapleLeafCap 0.42*

@hash_dd Personally think equilibrium will be set at this net 0 inflationary price

@hash_dd As for now it’s inflationary above 0.25; as ARR grows and as halvening continues the ceiling will keep getting raised

@zephyr_z9 159516 CH

Good point CXMT will generate over $130B in gross profit in the next 18 months How they spend that money will determine how long the memory cycle lasts and what happens in 28-29

$159516CH

Either that or you treat MSTR as a call option (once he cancels pref dividend) where strikes keep adjusting up.

Hmmm it's not looking good bruv someone need to bid this BTC now tbf MSTR will start accuring time premium when it gets closer to strike (~$25k) but the implied interest rate would be north of 15% per annum given dilution needed to fund pref dividends

再聊聊这波下跌的触发点: 从年初小龙虾这种个人 agent 开始引领的 tokenmaxxing 故事线的冷却吧. DeepSeek 被说是中国国运的来源应该是黑神话悟空的冯骥的微博, 而为什么 DS 份量这么重, 当时 2025/1/20 其实很多人都没有意识到它超出自身技术之外的, 对世界经济主线的影响, 我记得当时出来后很多人最大的乐趣是问它关于 "天安门广场" 的问题, 看着就很无语. 还是那句老话: 很多时代转折点出现的时候, 处于时代漩涡中心的人们大多数人只会认为自己度过了一个平凡的下午. 在今天市场开始 focus 在 "AI 头部公司还有没有钱烧"上面的时候, 回头看这一切的开始其实就是 2025/1/20 前后 DS R1 公开"蒸馏"技术的那个时间点. OAI 变得完全闭源其实也是从 O1 模型开始, 当他们内部发现"蒸馏"技术后, 应该立刻意识到了这意味着什么, 所以不惜毁掉自己 "Open AI" 的牌坊也要转向闭源, 如果这个技术一直没有公开流出, 甚至出来时间晚个1,2 年, 其实美国经济这边处理最难处理的债务问题会游刃有余很多. 最好的剧本是: 美国星际之门计划成功 (估计很多人已经忘记这玩意儿了) -> 头部科技公司 CAPEX + 政府资金 = AI 基建资金洪水 -> 投资带动就业 + 消费, 数据中心的建设带动相关制造业产业链回归 (芯片, 电力, 房地产, 半导体设备, 原材料 and etc) -> Opus 4.5 诞生, 进入 AI agent 时代 -> token 卖出高价, 全世界公司买单, 为什么? 因为你不买单你对手买单, 你对手的员工产出可以高你 10 倍. 人没办法一天写几千行代码, AI 可以轻松做到, 如果其他公司不焦虑, 华尔街可以用 AI 写一堆带着诱人 ROI 数字的研报让你们焦虑, 买单 -> 卖 token 开始带来大量盈利, token 经济开始正循环, 赚钱 -> 加大投入 -> 出现更好模型 -> 画更大的饼 -> 世界买单 -> 赚更多钱, 债务问题那时候提出来就只会让人想笑了. 而这个可行性很高的剧本没有实现, 到今天市场开始焦虑 token 投资回报率问题, 一切的起点只是因为 DS 公开了蒸馏技术, 而只要基础架构还是基于 transformer, 没有根本上的技术突破, 那蒸馏就是梦魇, 会一直缠绕着美国这边的头部 AI 模型公司. 那为什么中国要选择这种路线? DS 的开源如果说是梁文锋的个人选择, 后续的 qwen 开源, 到今天的 glm 开源权重, 这背后可以说是中国 ai 技术界的整体选择. 因为对中国来说, 只要需要硬刚工业基础的竞争都是舒适区, 开源模型权重, 那剩下的只是比拼和竞争对手的工业基础建设的成本问题了(小到一个变压器成本, 大到一条芯片生产线成本), 而这一点上美国相对中国的劣势是很大的, 而如果选择闭源, 等于是放着自己的枪不用去和对手冷兵器对砍, 当然不会这么做. 而且中国那边押注 AI 更大的押注是赌这波 LLM 代表的 AI 最强能力 -- 更高级别, 带自主决策能力的自动化 -- 普及后, 最受益的是自己最擅长的新兴制造业: 机器人, 新能源汽车, 各种智能穿戴设备 and etc.

@0xHoward_Peng @leopoldasch Don't understand your question sorry. I think TSMC not being on US soil puts it on a disadvantage. Seems more likely that the INTC x TSLA x SPCX x NVDA supercomplex is going to have a field-day in that scenario.

@MapleLeafCap @leopoldasch Interesting analogy; but that would mean TSMC might have an overvalued stock price and is also subject to significant "physical limitations." Correct?

During the peak years of WWII: - 50%+ of US market-cap was war-converted - 40% of US GDP was weapon spending (from 2% in 1940) As of today, semiconductor is - 18-20% of US stock market cap - 2-4% of US GDP in semi spending If you buy into @leopoldasch 's last chapter and that US is going to launch "The Projeect" in 1-2 year's time and this is a full-out war on intelligence. There won't really be a semi cycle and we are going a lot higher.

still here. locked in. forever growing. CT - my corner of the internet. get yours: s.kaito.ai/Wu3Jzne

@mikeahorton @brianzhusg @bidclubai I was looking back at it and u guys nailed 10 mm ARR. Burn / emission flip imminent. Huge congrats on the progress king!

@brianzhusg @bidclubai @MapleLeafCap we never left 😆

🔈BidCast - New Interview with @mikeahorton from @geodnet on $GEOD We discuss: 📌 Geodnet's device economics and growth flywheel 📌 Breakdown of its $3.5-4 mm ARR and path to $10 mm ARR with a diverse range of potential customers 📌 Tokenomics breakdown and math on net-burn 📌 Why Robotics is their #1 focus today with upside to $1 Bn ARR target longer-term Links 👇 Summary → tinyurl.com/5xr5zpdw YouTube → youtu.be/Rlddw6QICUY Timestamps 👇 00:00 Opening 00:44 Introduction to GeodNet and Its Technology 03:45 Understanding the Cost Differences in Base Stations 06:43 Market Competition and Hardware Evolution 09:34 Tokenomics and Network Coverage 11:19 Marketing Strategies and Community Engagement 13:14 Empowering Small Businesses with GeodNet 16:40 Manufacturing Challenges and Decentralization 18:46 Revenue Growth and Customer Base 23:48 Future Projections and Backlogs 27:45 Understanding Customer Dynamics and Token Economics 31:40 Exploring New Use Cases and Revenue Opportunities 35:23 The Role of Robotics in Future Growth 40:35 The Impact of Robotics on Market Dynamics 45:16 Innovations in Indoor Positioning and Satellite Technology 50:44 Competing in the Global Robotics Market 57:33 Financial Sustainability and Future Outlook

I personally think the $CRCL investors are significantly underestimating the risk of leading fintech players teaming together and launching their own stablecoin. For instance, the float on Stripe alone at any given point I think is around 15-25 Bn USD on its massive transactional volume. Already 20-25% of USDC without them actively juicing this on-platform float.

@hansolar21 Not kidding about Quadrillion market cap. It's got the best shot at it vs. any company ever existed.

@MapleLeafCap Gud bullpost on $spcx

Are you telling me your nitwit wallstreet brain is gonna short the 1st new empire since the united states and soviet union that's effectively modern VOC that will build factories on the moon and launch satellites into space at 0 marginal cost that will dominate world compute and be 1st to corner sun's energy via dyson sphere and mine the fuck out of asteroids this thing will be the 1st quadrillion market cap empire and outlast the USA and China?