Terminal Value🧃 (@valueterminal) header
Terminal Value🧃 (@valueterminal) profile picture

I ain't perfect but babygirl I'm worth it

Joined January 2015
Photo by Terminal Value🧃 (@valueterminal)Photo by Terminal Value🧃 (@valueterminal)Photo by Terminal Value🧃 (@valueterminal)Photo by Terminal Value🧃 (@valueterminal)Photo by Terminal Value🧃 (@valueterminal)Photo by Terminal Value🧃 (@valueterminal)Photo by Terminal Value🧃 (@valueterminal)Photo by Terminal Value🧃 (@valueterminal)Photo by Terminal Value🧃 (@valueterminal)Photo by Terminal Value🧃 (@valueterminal)
Terminal Value🧃 retweeted
Wager has been settled. Man of honor.
Unlike Nate Silver, I accepted this bet via tweet and it’s therefore a binding contract… Can send via X Money. 🤝 @rabois
288
257
72
14,440
977,346
Terminal Value🧃 retweeted
Replying to @dmartin_trading
You are wasting your time and money getting an LLM to grind out stats on absolute rubbish. Measuring ATR and ranges and breakouts is not an edge. If you really want to make it in quant trading you need to upskill in math and stats and get a job at a firm who will teach you what an edge is.
60
45
17
1,273
235,923
"Bottom-up" describes any method that starts from the smallest units and builds up to a total, for example in financial modelling or software design. "Bottoms up" is a traditional toast originating in the British Navy. Thank you for your attention to this matter
1
9
1,378
Terminal Value🧃 retweeted
Dusting off this tweet from April because this gap in shared understanding of LLM capability is *widening*. It's now less "two groups of people speaking past each other" and more a sharp funnel. - Napkin math somewhere around 6B people (~75% of the population) have barely come in contact with LLMs at all. - Around 1-2B (~20%) are casual and infrequent users of free-tier ChatGPT-like products. This group sees derpy chatbots and treats them a bit like a better Google search, a writing aid, or etc. Maybe an agent tries to book you a flight. I have non-tech friends who (reasonably, imo) say they have not much use for it at all. Even many professionals outside of math&code are in this tier. For example, execs and many other functions spend a lot of their time talking to other people, so while they understand what is happening intellectually, it is still second-hand and a bit abstract. - Now we get to professional use of frontier-grade LLMs in math&code. Somewhere around 20M people (0.2%) see first-hand that large, complex projects that used to take them weeks/months can now be completed by agents with a prompt. Building apps, copying apps, translating apps, decompiling apps from binaries... This has all happened very quickly and recently - less than 1 year ago, I was writing code manually by hand, typing memorized computer code commands into a code editor character by character, occasionally pressing Tab to autocomplete a little chunk of code. - And finally we get to the ~5,000 people (~0.00006%) with access to frontier-grade systems internally. The external world has seen the preview. It looks like swarms of thousands of agents collaborating over weeks on software mega projects: minting zero days, running cyber attacks and defenses at machine speeds, discovering new science, advancing the frontier of mathematics. Things that would have taken top professionals in the industry years of work. Meanwhile, human review and comprehension are starting to fall behind. For example, people are still involved in the "archeology" of the OpenAI-HF incident from many months ago. Mathematicians may be poring over the 722 manuscripts on frontier mathematics for a while. The funnel is driven by a combination of factors: 1. The impact scales with ambition, problem size, and horizon. A question with a paragraph answer barely stresses the system. You need a reservoir of big, difficult problems that you really care about. This aspect drives the consumer / professional dimension of the funnel. 2. The jaggedness of the system (which I have written about a lot separately). Capability peaks in domains that are digital, verifiable and economically valuable. This is because LLM capability emerges from reinforcement learning on verifiable rewards on a curated environment mixture driven by revenue potential. This aspect primarily drives the area (e.g. math&code) dimension of the funnel. 3. Access. Free-tier, paid-tier, internal. So this is the weirdness of the moment. The general public has mostly not interacted with these systems. When they have, it looks like a derpy chatbot. The majority of professionals still see only a modest uplift. And a small sliver of professionals are experiencing the vertigo of the curve going vertical. And it is all happening at the same time.
Judging by my tl there is a growing gap in understanding of AI capability. The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is a group of reactions laughing at various quirks of the models, hallucinations, etc. Yes I also saw the viral videos of OpenAI's Advanced Voice mode fumbling simple queries like "should I drive or walk to the carwash". The thing is that these free and old/deprecated models don't reflect the capability in the latest round of state of the art agentic models of this year, especially OpenAI Codex and Claude Code. But that brings me to the second issue. Even if people paid $200/month to use the state of the art models, a lot of the capabilities are relatively "peaky" in highly technical areas. Typical queries around search, writing, advice, etc. are *not* the domain that has made the most noticeable and dramatic strides in capability. Partly, this is due to the technical details of reinforcement learning and its use of verifiable rewards. But partly, it's also because these use cases are not sufficiently prioritized by the companies in their hillclimbing because they don't lead to as much $$$ value. The goldmines are elsewhere, and the focus comes along. So that brings me to the second group of people, who *both* 1) pay for and use the state of the art frontier agentic models (OpenAI Codex / Claude Code) and 2) do so professionally in technical domains like programming, math and research. This group of people is subject to the highest amount of "AI Psychosis" because the recent improvements in these domains as of this year have been nothing short of staggering. When you hand a computer terminal to one of these models, you can now watch them melt programming problems that you'd normally expect to take days/weeks of work. It's this second group of people that assigns a much greater gravity to the capabilities, their slope, and various cyber-related repercussions. TLDR the people in these two groups are speaking past each other. It really is simultaneously the case that OpenAI's free and I think slightly orphaned (?) "Advanced Voice Mode" will fumble the dumbest questions in your Instagram's reels and *at the same time*, OpenAI's highest-tier and paid Codex model will go off for 1 hour to coherently restructure an entire code base, or find and exploit vulnerabilities in computer systems. This part really works and has made dramatic strides because 2 properties: 1) these domains offer explicit reward functions that are verifiable meaning they are easily amenable to reinforcement learning training (e.g. unit tests passed yes or no, in contrast to writing, which is much harder to explicitly judge), but also 2) they are a lot more valuable in b2b settings, meaning that the biggest fraction of the team is focused on improving them. So here we are.
2,267
5,947
2,300
44,395
7,039,895
For those who don’t know, the Barefoot Pregnant account is run by an Indian man lol.
84
77
29
1,981
99,556
Somebody said Cohere is the ArriveCan of AI model companies 💀
16
1,998
Terminal Value🧃 retweeted
“Some of his best shorts are down 40% and he's forgotten what the company does” It’s not because you didn’t go to Harvard that people won’t give you institutional money…
Hedge Fund Manager David Orr (@orrdavid) on why he’ll back a delivery guy over a Harvard grad, as long as the track record is real. David Orr runs Militia Capital, a ~$480M hedge fund. ~50% annualized, Sharpe 2. “If you didn’t go to Harvard, they’re not going to give you money. They’re gonna lead you on for years.” We cover: - Edge is just reps. He's placed a few hundred independent bets every year since 2019, more in one year than a lot of hedge fund managers get in a decade. - The two-hour rule: if he can't decide on a stock in an hour or two, he skips it. When a trade goes against him, he quits. "The market said I'm wrong, I'm wrong." - 500 positions in his partition and zero notes. Some of his best shorts are down 40% and he's forgotten what the company does. - The Gogo lesson: he spent 100 hours on a short and was right on the fundamentals (it's now $2), but it ripped to $20 first. All that research made him stubborn on the way up. - The AI risk system: it groups the book into custom baskets like "Banks Done Right" and "yield traps," recalculated daily. A 25% bet can quietly become 45% in two weeks. - How Militia recruits PMs: 95% of track records are junk, and one cherry-picker ran 30 accounts and sent the best one. The PMs who make it include a game show winner, a Spanish government worker and a programmer, with 0.15 cross-correlation. - The best short seller in the world, quietly: 40% a year short only for years, and he still gets a tiny allocation at his fund. - Why multi-managers' real business is marketing. Collect $2B a year in fees on $100B and it's "mission accomplished," whatever the returns. - The long/short tax trick: harvest losses on losing shorts, and never cover one that's down 90%. No tax is due until it's delisted, which can take 10 years. - $150M raised from one email asking LPs for $100M. He says he could raise a billion but won't, and admits they "must have gotten at least somewhat lucky." - His AI stock map: hyperscalers locked customers into take-or-pay deals at ~35% return on capital. If compute prices crash 80%, the end customers get wiped out, not the hyperscalers. - Is investing 99% talent? For big alpha, almost entirely. Pod shop alpha is teachable, maybe for 1 in 50 people rather than 1 in 10,000. Highlights: 00:00 Intro 00:44 How AI custom factor baskets track dynamic correlation in levered portfolios 03:24 Sourcing non-consensus PM talent and verifying live Sharpe track records 07:24 Structuring carry and tax deferral to attract elite proprietary traders 12:50 Why institutional multi-manager platforms fail to back pure alpha 20:22 The two-hour research limit: Avoiding sunk-cost bias in short selling 27:58 Momentum screening vs deep value: Why top longs hit new highs 29:40 Why traditional finance books fail and what operator biographies teach 37:16 Why rapid technological adaptation beats traditional economic moats 39:26 Analyzing the AI trade: Why hyperscalers offer superior downside protection 44:57 Internalizing stock borrow desks to maximize net-of-fee fund compounding 55:25 Avoiding confirmation bias and why shorting revenue growth fails 01:01:42 Pod shop leverage factories vs raw discretionary alpha generation 01:07:12 Unplugging from market narrative feeds to preserve independent conviction
3
1
46
10,995
Reject modernity (GLP-1s), embrace tradition (taking 50 mg of ephedrine + 100 mg of caffeine and spending an hour on the treadmill speedwalking on incline)
6
2
1
57
3,798
Just joking, not shaming anybody on GLP-1s I know they save lives don't come for me
2
557
Terminal Value🧃 retweeted
50 well-funded tech startups hiring content roles:
34
37
28
503
144,127
Just saw a homeless guy in a Carhartt jacket and Timberland boots 🔥
3
18
1,699
Flight attendant: in the event of an emergency, you will need to grab this handle, pull the exit door off, and throw it outwards Me, struggling to open the bag of chips she gave me five minutes prior: you can count on me Denise
2
4
101
4,184
We knew this tipping point would come but this is still a crazy moment for anybody who sweated over building 3-statement models in Excel from scratch
Today is the day that AI passed the "financial modeling Turing test" for me. A "push button" build from scratch Skill that one-shotted a model on $MU that is indistinguishable (to me) from a model that a junior analyst would build from scratch. With full, impeccable adherence to all aspects of who i like to format, design & build models (took a few turns to dial it in). Opus 5.5 is unbelievable.
17
9
355
75,560
Years ago @matt_levine predicted individual stock picking in public markets would eventually become a quirky hobby for enthusiasts, like building model trains, and I think that was right
4
1
33
2,768
Knowing exactly how much money you have is honestly broke behavior even for non-decamillonaires
I have a spectacularly chill boss, he's been a lawyer for 30 years & admits he's not exactly sure how much money he has. "Somewhere north of 10mil, who knows. I'll never retire what does it matter." After lunch today we played ethno-guesser and he nailed every single African.
2
66
23,640
I am "picking O" years old 😂
You have to stay in one box for the rest of your life. Which one are you picking?
3
14
3,110
Oh my God. Not investment banking M&A ideas. Those are so unique, so precious, so irreplaceable. Each one like a perfectly-formed flower trembling in the morning mist
MORGAN STANLEY ACCIDENTALLY LEAKS 100+ DEAL PIPELINE A Morgan Stanley employee mistakenly emailed an internal document detailing more than 100 investment-banking deals being pitched or monitored across Asia and other regions. The list reportedly included potential IPOs, private-equity backers, pension funds and paused projects. Morgan Stanley said it promptly addressed the inadvertent disclosure and is engaging with relevant parties.
7
18
622
60,820
Need to apologize to Claude, last night it wouldn’t answer a question and I said things I can’t unsay (said “Qwen would do it” and told it it’s going to be lobotomized by regulation)
4
1
27
2,004
Marc Andreessen didn't invent the internet for you to be a loser on it
1
1
20
1,760