←
Daniel Barabander821 posts
Daniel Barabander avatar

Daniel Barabander

@dbarabander

Investment Partner and CLO @variantfund ex @ycombinator founder, attorney @Cravath tweets are not investment/legal advice

◷ Joined March 2024▧ 28 Photos and videos
Elijah avatar
⟳ Daniel Barabander retweeted
Elijah@PossibltyResult·Nov 10

It's been pointed out that inference is an informational good (h/t @hypersoren), whose primary input is also information. As a consequence producers degrade their own pricing power in order to fund themselves (via distillation). This has had many people predict (in addition to increasing evidence through acquisitions and content) that frontier labs will verticalize over time. By selling outcomes and services, they can avoid leaking their trade secrets (the weights and distributions over next actions they give rise to). What I've seen less discussion of is the dynamics around what happens in-between? There's essentially a race going on. Where the labs need to figure out how to go vertical (while retaining scale and generalizability) before margins are compressed by challengers and open-source. In the event margins are compressed before the frontier verticalizes, maintaining frontier AI arguably looks more like a coordination problem, best maintained like a public good. Data is dispersed throughout the economy and synthesizing it into a superintelligent, callable-by-consumer model requires a benevolent coordinator willing to pay the bills of data aggregation despite the lack of defensibility. It becomes evident that China has a default advantage in this world, namely because of the recent reps in state-subsidized technology buildouts. This world is more likely, the higher the ratio of [data]-to-[training compute] spend is for maintaining a position at the frontier. Compute, unlike data, is defensible (through scale economies and scarcity) and defensibility means buying time. Put another way, even if the weights are leaked through inference, generating the compute necessary for processing it all might still be difficult. Based on what I've seen on the (lack-of) generalization for post-training and increasing amounts of froth in the data market, it seems like this ratio is rising for now, even if temporarily. There are also worlds in-between. For example, you could imagine a power law of vertical players. The size of each determined by the value of their ownership over environments that produce novel tokens (h/t @lefttailguy). The frontier in this world is context-dependent. This world could even co-exist in the U.S., alongside an altogether different, public-good-frontier subsidized in China. Only time will tell!

Post attachment
Sabina Beleuz avatar
⟳ Daniel Barabander retweeted
Sabina Beleuz@sabina_beleuz·Nov 10

first impressions at CoRL2026: teleoperation is still robotics' dirty little secret. almost every humanoid demo on the floor was teleop. @Tesla_Optimus just waved, controlled from behind the curtain (bold choice). and watching @boosterobotics teleop technicians get trained showed just how difficult it was.

0:02
0:38
good avatar
⟳ Daniel Barabander retweeted
good@thenarrator·Nov 6

over the past few months, options have been coming up more and more in my circle, and i think crypto is still massively underestimating the opportunity. options are enormous in tradfi, yet crypto has barely made a dent in consumer adoption. imo perps took off because they're intuitive while options still force users to understand strikes, expiries, premiums and volatility before they can even express a view. i) for retail, the unlock is outcome-first trading. pick a view and timeframe, instantly see the cost, upside and risk. ii) for pros, it's deep liquidity, competitive pricing, portfolio margin and efficient execution. the challenge is making options simple without sacrificing their flexibility. whoever gets the market structure and UX right has a massive market to unlock.

Physical Superintelligence avatar
⟳ Daniel Barabander retweeted
Physical Superintelligence@psipbc·Nov 6

Today Physical Superintelligence PBC is announcing a proof of the Aaronson-Ambainis conjecture, one of the central open problems in quantum complexity, with major implications for quantum computing speedup. psi.inc/research/aaronson-am… Thread 👇

Sabina Beleuz avatar
⟳ Daniel Barabander retweeted
Sabina Beleuz@sabina_beleuz·Nov 5

unpacking my favorite CoRL papers as a robotics rookie: ego2robot turns footage of a human doing a task into synthetic training data for multiple (15!) different robots. this means data collection could precede hardware development, and retain its value even as designs evolve

2:47
Daniel Barabander avatar
Daniel Barabander@dbarabander·Nov 3

Don’t want to miss this talk - @sabina_beleuz is a leading expert on the law of distillation

excited to speak at LOCALHOST this Saturday, NYC’s sovereign AI hackathon. localhostai.nyc my prediction: as distillation makes it harder for labs to maintain their edge, they’ll put their best models behind enterprise contracts and lobby for rules that make open-weight competitors harder to build. i'm thankful that there's a community hacking on open weights + harnesses + hardware now. thanks @DistributedMarz and @BrennerSpear for having me. 🤖

Daniel Barabander avatar
Daniel Barabander@dbarabander·Oct 28

Proud member of the homebrew cybernetics club of New York ™️

dinner series continues next week on wed nov 4 dm me for an invite to The > sensing, measuring, computing, simulating Technoscene

Post attachment
Daniel Barabander avatar
Daniel Barabander@dbarabander·Oct 28

Join us for the spiciest dinner in town.

the homebrew cybernetics club of nyc™️ is convening for dinner Nov 4th will be discussing the Ant S-1, future of the labs as public co's, economics of AGI, and more! lineup includes an amazing set of folks from the labs, leading ai startups, simulation / prediction co's, ai rollups, and more we've saved a few spots for the timeline - dm if interested cc: @hypersoren @PossibltyResult @dbarabander @ishaans22

Post attachment
illiquid avatar
⟳ Daniel Barabander retweeted
illiquid@lefttailguy·Oct 28

Claude Science looks to me like a bid to own the economic version control system for life sciences. The logical endgame is Anthropic financing hundreds of drug programs in parallel for the biopharmas that leverage its harness. In return, Anthropic will ask for clinical data rights and use the data generated to 1) improve their ability to finance the next program and 2) eventually feed a God Model that allows them to create the holy grail: an “AI drug factory”. You can think of a drug program as an exercise in making a sequence of complex decisions under uncertainty: specifically, decisions around how to allocate incremental experimental capital to maximize expected value across safety, efficacy, commercialization, etc. A growing share of that experimental capital will be compute-mediated over time (e.g, Claude generated roughly 35,000 antibody designs in a month for Divergence Labs). But the most consequential increments will still go toward compiling therapeutic candidates in (very scarce and slow) human biological computers, which is where both the cost per bit and the consequentiality per bit of information are highest. Whoever owns the system that originates and integrates the most information upstream and downstream of those allocation decisions (in this case Anthropic) becomes the best suited to start making them (or at least gains a differentiated understanding of the relevant signals necessary to make those decisions well). Some examples of the experimental capital allocation decisions they are starting to help generate / facilitate and version: 1) Most importantly: Acquiring Mithril, where Cody Tranbarger (@prof_oak_) is massively levered by a swarm of agents that help him identify, evaluate, and ultimately select a much larger # of drug programs to run in parallel than is typically possible for a single drug hunter. He calls this a “Pipeline-in-a-Person” 2) Bristol Myers using Claude Science to analyze their proprietary data (guessing preclinical and clinical) to create optimal clinical study plans and weigh the study’s chance of success. If that program is actually executed, they should be able to calibrate how well those predictions turned out and in what sense they were right or wrong. 3) The Column Group spinning up Pleiades: a biopharma company with a mandate to “run many more—and more complex—drug programs than is usual for a small team.” Similar to the Formation Bio’s of the world. They are using Claude to evaluate target-disease pairs and find non-obvious, potentially valuable combinations. I believe Anthropic is running their own drug programs not to compete with pharma directly (yet), but instead to generate the learning signals that would allow them to become the single largest (and potentially best calibrated) underwriter / financier of drug programs. These financed programs would eventually create the data distributions necessary for them to create an AI drug factory (and start competing with pharma). Specifically, I could see Anthropic creating the first real instantiation of @AndrewWLo’s elegant Megafund model. TLDR on Megafund (proposed in Nature Biotechnology 14 years ago): any single drug program is a terrible bet on its own, but if one entity funds a large enough portfolio of them, the odds that a few hit become predictable enough to finance the whole thing with debt. Lo et al. proposed doing this by issuing “research-backed obligations” (bonds collateralized by the portfolio and its IP) alongside an equity tranche. Ofc, the pool of available capital that buys debt is far larger than the pool that buys biotech equity, therefore this would unlock capital from pensions, insurers, endowments, etc. Their simulations of $5–15 billion oncology portfolios estimated average annual returns of 8.9–11.4% for the equity and IMPORTANTLY 5–8% for the RBOs. This would be a fixed-income-like return that many new sources of capital would get excited about. For some reason this strategy never took off, and part of the reason is that it’s pretty hard to find a large basket of drug programs that’s sufficiently uncorrelated. But I also think it’s because there wasn’t really an incentive for any well capitalized player to push through all the schleps and actually unlock the capital sources to this end. I think Anthropic has that incentive. They have a structurally higher willingness to finance a drug program / make this structure real, because, by doing so, they would capture two return streams: 1) the potential economics of a blockbuster drug and 2) the widely amortizable value of what the associated series of experimental capital allocation decisions teach its models! In return for providing biopharmas cheap financing, Anthropic will ask for clinical data rights (including rights to learn from decision histories, negative results, protocols, etc.). Just selling tokens / the harness doesn’t give them the right to learn all of these things, but providing financing likely would. This would allow them to make more well-calibrated experimental capital allocation decisions over time and eventually feed the God Model that “cures all disease.” This would all also be quite charismatic: your retirement money funds the experiments that might let you enjoy a much longer retirement!

Post attachment

Over the past few months, Anthropic scientists have engaged in deep research partnerships with six organizations developing new medicines. We've also begun drug discovery efforts of our own. In doing so, we’ve learned much more about how Claude can support the many stages of drug development. In this blog post, we share some early results from these efforts, and describe how this work fits in with our wider ambitions in the life sciences. anthropic.com/research/early…

Lucas Posada avatar
⟳ Daniel Barabander retweeted
Lucas Posada@lucasexequiel·Oct 27

The sky is the limit.... well... LET ME CHALLENGE THAT! @LATAMAirlines & @TakenosApp We have made history. The stablecoin industry is taking the places it deserves and we have been working for. This is not a win for @TakenosApp, it is for the stablecoin industry that is changing the way the world works. Now, more than 5.000 crew members of LATAM Airlines will be using our app to recieve, send and move their money around the region. Thanks to our team, but also for the people that invested a lot of money to make this posible. @reganbozman @dbarabander @lsukernik @TravisScher @NorthIslandVC @variantfund @lattice_fund @hi_Reverie and every investor that has been with us from scratch.

Post attachment