HF PM G10 / EM Rates
Singapore
Joined November 2016
- Tweets5.5K
- Following1.2K
- Followers249
- Likes861
Iron Nuts retweeted
This is just too cool.
Surprise 1: Anthropic acquired a one-person biotech
.
.
.
Surprise 2: that person is me
Surprise 3: I’m here to tell you all about it
Yes, this is what you think it is; after 6 yrs of increasingly loose anonymity: the Professor Oak reveal post
…and yes, it has gotten so boring around here that I’m literally doxxing myself to spice up the timeline
-------
Ok, allow me to introduce myself. My name is Cody Tranbarger. I mostly tweet, but occasionally find time to dabble in drug dev & starting/building biotechs. My past lives span the VC, crossover, startup continuum; Prof Oak has been my way of keeping a digital foot in those worlds, despite having my physical feet planted elsewhere.
The past 18 mo, ‘elsewhere’ was Mithril Medicine, a co. I founded to develop bsAbs for LAMA2-CMD & other ECM-related genetic diseases.
Mithril was born out of hands-on building with @MikeNGladstone, @JasonPaulRhodes, @LifeSciVC, & team @atlasventure. Somehow, I made enough half-decent calls along the way (even a broken clock is right twice a day, eh?) to convince them Mithril/I warranted a small seed. While I can hardly rationalize such an outrageously cavalier bet, I’m grateful for it, & I'd partner with Atlas again in a heartbeat.
In typical AV fashion, Mithril’s primordial form was virtual, decentralized, & ultra-lean. This stage tends to be short-lived, but for Mithril, it worked surprisingly well, for surprisingly long, thanks in part to the rapid maturation of the AI tools I used liberally.
Eventually, Mithril became a bit of an experiment itself - to find the upper bound of how far it could scale as a single operator + a bunch of Claude (+ a stellar bench of part-time consultants).
In parallel, I started spending time with @nc_frey, @an1lam, & @CoefficientBio team; and promptly experienced, for the 1st time, AI at the frontier. Soon thereafter, Mithril stopped looking like an experiment, & started looking like a preview of an entirely new operating model.
From that spark came “Pipeline-in-a-Person.” The idea immediately resonated when we shared it here back in April; and it was central to the shared vision that led to Anthropic’s acquisition of Mithril, & my joining the team.
Why join Anthropic? Three reasons.
(1) For Mithril, the programs continue, with more resources & infrastructure behind them, and as much Claude as I can put to work.
(2) For me, the chance to start several more programs, & find out how many one person can actually run in parallel - an almost unprecedented luxury for a curious drug hunter.
(3) For our mission, a path to institutionalize & scale this new way of working, then put it in the hands of all of biopharma, all over the world - which, candidly, I find the most energizing.
Will AI x Bio cure diseases? Unequivocally yes, imo… but not by the route that gets the most airtime. Rather than timeline acceleration & compression, I see orchestration & parallelization as today's low-hanging fruit.
Most drug hunters see a handful of go/no-go calls in a career, & spend entirely too much of the rest trading emails (my inbox can attest).
Give them the bandwidth to run several programs at once, and the time to focus on scientific judgment, and I’m willing to bet they’ll make better calls. And at scale, better calls translate to more drugs, higher hit rates, broader patient impact, & likely, a disease or 10 or 100 cured. I couldn't be more excited to build toward that future @AnthropicAI.
Finally, a heartfelt thank you to BioTwitter/X. The end of anonymity is bittersweet, but if anyone’s earned a proper introduction, you all have.
It’s difficult to overstate how much this app has done for me: a decade of education, plenty of friends, many job opportunities, several mentors, a few investors, even a co-founder (fmr). It was the generosity of this community that inspired me to start posting 6 years ago, after just as many years lurking. I hope I’ve been able to pay it forward, in my own small way.
But, to be clear, Prof Oak isn’t going anywhere. Same beat, same themes, and now, a front-row seat to drug development at the frontier. Stay tuned for dispatches from the trenches; and remember...
Gotta Cure 'Em All
Iron Nuts retweeted
Don, there is a missing term in your equation.
The crunch spread is not: Token Price - GPU rental price.
Rather it is: (Token Price * Token Volume) - GPU rental price.
More simply: Revenue per GPU - GPU rental price.
Continuous throughput improvements for the same exact token mean that volume growth can more than offset token price declines for the exact same model/token
i.e. Kimi K3 token throughput per GPU has recently been improving at a 37% annualized rate driven by inference engine optimizations.
GPU prices are vertical because:
1) Mix is shifting from higher margin frontier tokens to lower margin open-weight tokens. $10 in Claude revenue requires $1 of inference compute. $10 of of open-weight revenue requires $3 to $7 of inference compute given competition between inference providers leads to lower margins. Open-weight tokens taking share is positive for AI infrastructure demand given lower margins at the model layer. The reverse is also true in that frontier tokens taking share is negative for AI infrastructure demand. Open-weight models have been taking share since early summer.
2) Throughput is steadily improving for each individual token type such that GPU compute becomes more valuable even as true like for like token prices decline.
Iron Nuts retweeted
Wow. amazing tool.
The future of investment research is very clearly narrative-based or investigative (ie, Gavekal, Citrini, the Perkini thing).
Macro forecasting and analysis is nearly fully commoditized. But the world a year from now is always vastly different than we expect.
Policymakers remain human (for now), exogenous shocks (like covid) crop up out of nowhere, elections are always uncertain and increasingly consequential.
Edge is in being creative in a thoughtful way. Or flying to the factory and asking questions. Very different skillset than what dominated the past 10-15 years.
Iron Nuts retweeted
This increase in credit spreads is consistent directionally with what is predicted to happen to credit spreads based on Kim and Wright’s model estimates (half of the 10 year ust increase was term premium)
Of course if you zoom out from this picture, the latest increase is but a blip.
At the same time, my back of envelope estimation using the KM term premium estimates would say HY OAS can go up by another 100-140 bps soon.
Iron Nuts retweeted
The unconscious and intuition are fascinating to me, because it seems as if the conscious mind is operating on a much smaller subset of your total available intelligence
Which means if you can find a way to unlock more of its wisdom, it has the capacity to drastically change your life
What most ppl don’t seem to realize is that all of your experiences are being compressed into language and narratives in your mind that allow you to consciously reason with them. BUT that representation is incomplete. You actually have an enormous amount of information stored across your experiences that you can’t necessarily articulate, but that can still influence your perception and decisions
I like to this of this akin to an LLM operating within a narrow context window. If you’re talking to someone about the stock market, your thinking is naturally constrained by that topic, and you’re less likely to pull in something you learned from an entirely different domain that might actually be relevant. The information STILL exists, but your conscious attention isn’t accessing it (my hunch is that it’s energetically costly to context switch, but that’s a topic for another day)
Intuition seems to be one way that this broader intelligence expresses itself tho. Your brain can integrate patterns across experiences without you having to consciously retrieve and reason through each one. The result is a feeling or a sense that something is right or wrong, even when you can’t fully explain why via the ‘rational’ mind and language
You see this in elite athletes all the time, chess masters, the best traders, etc… When a hockey goalie like the guy below has milliseconds to react to a puck, there’s no time to consciously calculate the optimal response. The thinking mind is simply too slow. The athlete has to trust a much more deeply integrated form of pattern recognition that has been developed through thousands of hours of experience
And I think this generalizes far beyond sports too. There’s a time to slow down and think deliberately to plan a general path (think: technique to play goalie position), but there’s also a time when continued analysis actually interferes with your ability to act on what you already know (game)
I often wonder if we’ve all become so identified with the conscious mind in this day and age of “information abundance” that we mistake its limited, verbal representation of reality for the totality of our intelligence
If intuition can truly integrate more of your experiences than you can consciously articulate, then learning to access it is quite literally the single greatest thing you can do for your life (and everyone around you too)
Iron Nuts retweeted
If citadel is paying more and more attention to “factor & rotation shocks” so should you
Btw. Justin is a stud
citadel's global equities head says some of the biggest hedge-fund losses now happen without a major market selloff.
“The S&P 500 didn't move very much.”
justin lubell describes huge P&L swings driven by factor shocks and rotations inside the market.
a portfolio can look diversified by ticker and still be one crowded bet.
Iron Nuts retweeted
I think this recent volatility continues to reflect the misalignment in the product cycle and the capital cycle. The product cycle itself has gotten more complex not simpler due to the continued presence of multiple competing products and the shifting relevance of open and closed models. The capital cycle is adding complexity on the other side of complexity proving two things can be true. The market can be structurally under supplied compute/infra but the capital complexity to finance the build may be over extended and overly complicated. I observe a lot of frustration about this dynamic. If you are super bullish long term this should be great. I think the issue is people are more bullish about the economics caused by the under supply than perhaps their bullishness on the underlying unit economics of the products on the other side of the shortages. The frustration is driven by not getting/keeping a sustained multiple now on the current economics reflexively driven by the supply/demand imbalance. I wont pretend to know what’s going to happen to alleviate supply or not, or what the economics of the models will look like depending on the slope of supply. I just know this kind of uncertainty usually leads to volatility because the market hates uncertainty more than anything. That means being nimble. Just when you think there is a clear trend, something else will change and make you re-think everything you believe today.
The equity market has had a fifteen year tailwind of declining capital intensity, lower cost of debt, and higher margins fueling a debt and fcf driven equity shrink support to eps and multiples.
There is a generation of investors now that have not had to think much about capital intensity, capital cycles, return on capital, tightening liquidity.
Everyone is in agreement that AI is the most important thing. And AI can dramatically dwarf the capital cycle but the point is that this isn’t just about AI it’s AI vs the capital cycle. And the capital cycle can lose and the market can be fine. But if the capital cycle wins the equity market resets. The right tail is limited and somewhat reflexive based on whether AI would be purely additive or substitutive. The belly of the distribution curve of outcomes seems normal and the left tail seems increasingly fatter.
Iron Nuts retweeted
On Friday evening in London, the price of oil for most of the world moved almost 10% in minutes.
Nothing major happened in the Middle East that afternoon. No new macro data. Just a U.S. government announcement timed right before the key physical oil window closed.
Most people have no idea how oil prices are actually set. Some prefer it that way.
Not me..
On Friday (Oct 16), around 3:15pm London time, the U.S. DOE offered 17 million barrels from the Strategic Petroleum Reserve as an exchange. The offer itself wasn’t shocking. It was the remaining piece of the 172 million barrels the U.S. committed under the March IEA deal. Earlier big offers months earlier in 2026 had seen almost no take-up because of the high return premiums required.
What was unusual this last Friday was the timing.
The SPR announcement landed just as the Platts London physical Brent window was about to start. That window concentrates activity in a short period (roughly 4-4:30pm London). Liquidity is often thin on a Friday afternoon.
Oil market participants had almost no time to digest the news before books were marked to the assessment that sets a key global physical benchmark.
Prior large SPR offers months earlier had typically come earlier in the day with multi-day bid deadlines.
This one arrived late, right before the Friday evening close.
The effect was immediate in the physical market. Dated Brent printed sharply lower (down over $10 on the day). The physical differential collapsed. Prompt scarcity premiums were sold. Related futures shorts later covered as paper (futures try to track physical cash prices) moved the other way.
The result: one of the largest one-day physical-cash price down / paper-oil (futures) up divergences on record.
The physical oil market is concentrated. A small number of players in London, trading a thin set of cargoes, can move the assessment that prices a huge share of the world’s crude. A well-timed headline can whip that group around.
Governments regularly complain about oil price “manipulation.” They are quieter when the move comes from their own reserve announcement landing at the exact moment it can have the biggest effect on the physical benchmark.
Whether the Friday afternoon timing was deliberate or just when the paperwork finished is not stated in official sources.
What is clear is that it worked as a real-time catalyst in a thin Friday evening oil market.
How oil prices are set is more mechanical, and more sensitive to a handful of actors and announcements, than most people realize.
Is global oil a fair and free market when it has a very specific closing time (and clear headline target risk) and a tiny list of 5–6 people who effectively get to vote on what 8 billion people pay ?
Iron Nuts retweeted
By @Citi's math, we’ve been in a bubble since mid last year, and Dirk Willer, Head of Global Macro Asset Allocation, says to keep buying.
We also cover:
• Why 1929 is the only bubble that crashed immediately
• The Warsh shadow rate and why it tracked Greenspan’s Fed
• Why the bond selloff is about growth and not deficits
• Why Brazil’s 9% to 10% real rates set up a big opportunity
• Why gold peaks when the PBOC is done buying
• His favorite bubble hedge
Iron Nuts retweeted
I Implemented @karpathy's LLM knowledge base for stock research back in April. Six months later, I still use it every day. Here is how I have implemented it for stocks, and why it's a great fit:
treasure-hunting.co/p/how-i-…
Iron Nuts retweeted
According to new analysis asset managers have increasingly found the cash-futures basis trade less attractive relative to alternatives, including swaps linked to new debt issuance tied to data center construction in support of artificial intelligence (AI). dallasfed.org/research/econo…
Iron Nuts retweeted
Wow. I woke up to an impressive number of upset people who apparently don’t understand how LLMs work and can’t cope well with me pointing it out.
An LLM is a statistical next-token prediction model. That’s how it works. It’s an extraordinarily sophisticated calculator performing math you could theoretically do by hand (if you were sufficiently masochistic).
Math doesn’t know what it’s calculating. Pi doesn’t spend its evenings contemplating the nature of circles, and c doesn’t ponder the relationship between matter and energy.
That doesn’t make the results any less accurate, impressive, or useful.
Between inference requests, an LLM isn’t sitting around thinking. It isn’t reconsidering its answers, wondering whether it made a mistake, contemplating its existence, or questioning whether it’s conscious. Its weights remain fixed (it doesn’t learn), and computation resumes when it receives input.
We know this because we designed these systems. The architectures, mathematics, implementations, and research are extensively documented. You can study them yourself.
Of course, some people immediately lean into comparisons of biological intelligence or philosophical debates over consciousness because that is what they know.
But biological intelligence and artificial neural networks are fundamentally different systems. Similar outputs don’t establish equivalent internal experiences.
I experience consciousness directly. I infer that other humans do too because we share biological structures and processes. We don’t have equivalent evidence for an LLM, and nothing about its known operation gives evidence of conscious experience.
Which brings me back to my original point.
To me, understanding is conscious comprehension. It means being aware of information, recognizing its meaning, and experiencing that comprehension. If I wasn’t conscious I could not understand.
We can perform actions unconsciously. We can produce correct results without knowing why they’re correct. But I distinguish those capabilities from consciously understanding what we’re doing.
An LLM can solve complex problems, explain difficult concepts, and produce remarkably accurate information. I give it full credit for those capabilities.
But producing an answer that we rationalize as understanding isn’t the same as experiencing understanding.
You don’t have to agree with my definitions. We can debate functional understanding versus conscious comprehension all day, but that won’t establish whether an LLM actually experiences anything.
The right debate is this…
Could machines exhibit something different from human consciousness, something measurable, meaningful, and worth studying?
I think so, absolutely!
I’m entirely open to discussing “machine consciousness” in terms of observable, testable computational properties. We built these systems, we can instrument them, examine their internal mechanisms, and investigate what they’re capable of.
That seems far more productive than endlessly arguing over whether a machine experiences life the way humans do.
So here’s my question:
How would you define machine consciousness, and more importantly, how would you measure it?
Iron Nuts retweeted
Dated Brent DUMPING!!!!!!! hahahaha
Physic nerds crying about Nov Nsea programs or collapsing freight rates(arb Midland cargoes into Europe), who cares???????
And Dated-M1 still +$20? Don't care too 🤪 Good karma says -$2 is printing any day now!!!
$12 WHALE OIL IS INCOMING, OIL BULLS ARE SO COOKED RN!!!! I LOVE OIL GLUT ❤️❤️❤️❤️
THANK YOU FOR YOUR ATTENTION TO THIS MATTER!!!!
#oott #iran
