Finflexity

“Today's economic summary: -The tungsten crisis in Europe and America, a key element in weapons, China possesses 50% of the reserves and knows this. -A crisis in raw materials for plastics, due to the war and China's ban on their export. -Ammonia crisis and the closure of one of the world's largest ammonia plants in Australia. -Gas crisis in India.”

Thoughts from my Shanghai field trip Sentiment among Greater China investors was far worse than I expected. Unlike the US market, A shares had not seen a single rebound since July. Still holding Chinese and Hong Kong AI stocks that had crashed and Korean memory stocks that had collapsed, investors kept pressing me on when memory would bounce back. Many investors said the AI trade was over, and a few said they doubted whether even 2027 capex, not just 2028, would go ahead as planned. Overall, the bearish sentiment was so strong that even I, a bull before coming to China, started to waver. It got to the point where, faced with the bear case that memory prices could start falling in the third quarter of next year, I took it seriously and couldn't refute it. As the conversations went on, I gradually began to understand where their pessimism was coming from. In the end, they were all asking the same question. So what comes after AI coding? What will be the next engine of ARR for Anthropic and OpenAI? Several candidates came up. AI4S, AI chip design, and so on... But none of them clearly convinced investors. I couldn't offer a clear answer either. Sure, there are rumors that Anthropic has made major breakthroughs in both AI4S and AI chip design, but what we actually want to know is when any of that turns into money. With AI4S, there could be beneficiaries right away, such as companies that sell lab monkeys or diagnostic kits. In fact, there already are. But while we can trade those names in the short term, what we ultimately want to know is whether Anthropic can make money by developing drugs. That is still several years away. As for AI chip design, the TAM itself is hard to calculate, and even if you do calculate it, it is hard to monetize. That makes it difficult for the market to believe in. Honestly, this feels like the hardest market in history. If a company isn't making money, selling makes sense. But right now the market sells even when companies are making money, even when they beat, and even when they say demand will stay solid all the way through 2028 and 2029. So my recommendation to investors was to buy memory, and within memory, DRAM. Not because I have conviction in the DRAM cycle, but because I see it as a relatively defensive choice within the AI trade. Memory stocks are now trading close to 1x P/B on 2027 BPS. Historically, memory stocks have staged technical rebounds when they fell below 1x P/B. And even if the broader market falls further from here, the three DRAM makers will keep buying back their own shares. I thought that would guarantee a certain level of buying and concluded the stocks would have downside support. Of course, I don't know whether memory stocks can break above their previous highs. That depends on whether sentiment across the broader market recovers. But I think there is a fairly good chance that memory outperforms the rest of the AI trade.

Well worth reading: proofsandprompts.com/2026/10…

Brett Adcock is one of humanoid robotics’ biggest hype men. @Figure_robot is valued at $39B; @Forbes says he's worth $23.4B. Nearly two years ago, he said he saw a path to 100,000 robots across two customers in four years. How far has Figure gotten? Our profile 🤖: forbes.com/sites/annatong/20…

This slide from Coatue has to be one of the worst calls of all time Four year later, missed 4 / 4 and called Robinhood a broken business model with no valuation support after the stock was down 90%....

yesterday i was reminded that the world i experience is very different to what's actually happening for ~99% of the population. my girlfriend was working on her laptop and showing me her ai use, which is gemini because it's bundled with google business accounts. everyone at her company uses it and mostly the flash models because cheap and fast my guess is this is most people's experience with ai, that and chatgpt via the iphone app or web most still have no idea about agents, harnesses, mcp. tech and X is very much its own wee world

Kind of hilarious that we spent a trillion dollars trying to build God, only to realize that a cheaper version of this God is more than sufficient to complete 95% of the tasks in corporate America, who would ultimately be the end customer for this God anyways Turns out that you don’t need that much intelligence to hold up even the most powerful and vibrant economy in history Intelligence was never the bottleneck

BTW both Fable 5.1 and GPT 6 failed, overnight, on a problem that I solved in 5 minutes. I want to post what happened just to document why we're not quite on AGI yet, although getting closer every day. This post will be a bit technical because details matter So, last night Fable identified a memory leak on Bend2. A given long-running process would steadily grow in memory, until it crashed in just a few hours. That's the bug. I asked both Fable and Astra to fix it overnight. Rules: 1. Do not increase the codebase size 2. Do not cause regressions in any benchmarks Both failed. But why? First, both identified the correct root cause. (everything below is written by Fable) Bend2's runtime has 16k GPU lanes. Each lane has its own private free lists, so alloc and free are 2 stores, no atomics. That's what makes it fast. The catch: when a lane frees a block that another lane allocated, the block lands on the freeing lane's list. If that lane never allocates as much as it frees, the block sits there forever. Nothing is lost, the memory is just parked on a list nobody uses. Both models found this. Both measured it: 40-85% of all frees cross lanes. Then both spent the whole night trying to fix the allocator. Return blocks to the owner lane: +124% slower. Slabs: 73x slower. Buddy: +7-21%. Flush at sync points: +3.5k tokens. A full allocator rewrite with lock-free chains: leak gone, but +1.9k tokens and matmul 78% slower on CUDA. Every design either broke the token cap or a benchmark. Neither model converged. They just kept building more machinery on the same premise. The premise was wrong. I asked for a histogram of where the parked memory actually was. 96% of it sat in the last 4096 lanes. That's not an allocator problem. That's the scheduler: it hands out work as a pure function of lane id. Lane 0 always gets the root of the tree. The same lanes always get the leaves. The join always runs on the last child to arrive, and the same lanes always arrive last. So every launch, memory flows in the same direction, from the same lanes to the same lanes, and the flow never reverses. The allocator was fine. The lanes weren't interchangeable. Fix: rotate lane identities. Each launch, physical lane p uses the free lists of lane (p + r) mod 16k, with r advancing by a golden-ratio step. Over a cycle every identity plays every role once, and the imbalance sums to zero. No allocator change, no atomics, nothing on the hot path. A handful of lines. Once I told Fable this, it implemented it, measured it, and ran the whole gate in an afternoon: 555MB → 23MB after 2000 launches, OOM pushed from ~7k launches to ~75M, zero cost on all 16 benchmarks. So what was missing? Not capability. Both models can write a lock-free allocator overnight. What neither did was step back and ask why memory flowed in one direction at all. They accepted the asymmetry as a fact of the system and built increasingly complex machinery to undo it, instead of removing it. That's the gap: not "can it do the work", but "does it question the frame it's working in". Still, it did everything else, from the diagnosis to the census to the fix, once pointed. Closer than last year. Not there yet.

Terrence Tao’s reaction to the potential news of a frontier AI model solving a hard mathematical proof problem is potentially also very instructive for the question of ROI from AI. A problem that’s very valuable for a human to solve with human reasoning that comes with the attendant insights and motivation for future problem-solving may not be as valuable if it were solved by a machine, esp the value of the technical solution of the problem doesn’t entirely lie in the destination as much as the process. An Anthropic model may well have solved the Navier-Strokes problem as posed on the Clay challenges acc to @AndrewCurran_. But it’s far from clear that it has high economic returns except in demonstrating the raw power in solving a hard unbounded task. At some point rather than being satisfied with being given hints or analogies for the power to do extraordinarily valuable creative things, people may start demanding to see the models to just do extraordinarily valuable creative things already.

The stats beyond the final score: If you are wondering whether Argentina surpassed Spain in any of the underlying statistics, it wasn't in shots on goal, possession, corners, passes, or anything positive. It was in the fouls they committed and the yellow cards they received. And of course the red card, but then that keeps their lead on this stat for all world cups! #worldcup #argentina #ArgentinaVsEspaña #spain @FIFAWorldCup

Let this graph sink in. Anthropic outpaced OpenAI in ARR. Now everyone understands why OpenAI is so keen on business and enterprise. And now everyone understands why Anthropic looks significantly better in light of its IPOs.

WSJ obtained confidential financials from both OpenAI and Anthropic ahead of their expected IPOs later this year. The core tension: revenue is exploding, but training costs are exploding faster. OpenAI projects $121 billion in compute spending by 2028, resulting in $85 billion in losses that year alone, even after nearly doubling revenue. Strip out training costs and both companies are near profitability now; add them back and OpenAI doesn't break even until the 2030s. Anthropic expects to get there sooner. Inference costs still eat over half of revenue at both labs, though that share is declining.

In September Applied Research Associates announced that they were awarded the development contract for NGP. Pentagon has said nothing. $194 million is budgeted across FY25 & FY26, plus additional funds for related R&D.