I write stuff sometimes
Joined January 2023
- Tweets17.4K
- Following2.1K
- Followers5.8K
- Likes36.6K
Happy retweeted
not only am i not worried about the crypto community's ability to handle the upcoming ai/quantum threat, i will venture to say that in the next 2 yrs the world will realize crypto is far better prepared for it than the rest of the internet and finance. we've had a trillion dollar *irreversible* honey pot for a decade, producing a far more battletested system and adversarial security culture than pretty much anywhere else on earth.
Happy 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.
Happy retweeted
Bitcoin's proof of work has done around 2^96 hashes over all time. Breaking the ZKP's in Zcash requires a minimum of 2^120 elliptic curve operations assuming ECDLP. If you had a dedicated ASIC, an elliptic curve addition would cost around 10-100x more than SHA2.
So today breaking an Elliptic curve would cost around 1 billion Bitcoin-hashrate equivalents. AKA a billion times the hardware buildout of all Bitcoin mining ASICs, dedicated to break this.
If a hypothetical billion X improvement came in, you'd need a chip buildout on the level of bitcoin mining before you can break anything. While competing against all the hyperscalars for fab-space.
AKA plenty of warning time, for a recovery process that takes a couple day response time max.
Happy retweeted
The ability to resist panic and think in dimensions rather than memes seems increasingly important in the age of AI
People read one emotional Matt Green tweet about public key encryption, spent 0 seconds thinking about it, and decide to panic.
Sure ECDLP will be broken with QC's one day. But there's so many levels to this being wrong to panic over.
- 0 actual progress on attacking anything
- We have tons of alternate plans available.
- Quadratic advantage hash-based PKE
- Lattices
- Multivariate polynomials
- A whole field of KEM's from fine grained cryptanalysis. (Fun fact, works even if P=NP)
Sure you can wake-up and choose doomer and be wrong.
We have decades of cryptography to draw from on for all the above options, AND how to build a lot of what we want from UX tweaks without PKE. PKE just made a lot much easier, and less web of trust routing like.
Happy retweeted
Claude has discovered a previously unknown enzyme system hidden in the DNA of bacteriophages. Beside the enzyme’s gene sits a long array of repeating DNA—a structure that looks somewhat similar to CRISPR.
We don’t yet understand what this system does, but only a handful of known systems share its features, and all of them are able to cut, copy, and paste DNA. Historically, the discovery of such programmable systems has helped revolutionize medicine. CRISPR, for instance, is now the foundation of genetic medicines. But it will take much more work to learn what this system does, and whether it can be put to similar use.
Read more: anthropic.com/news/claude-di…
Happy retweeted
zcash is simultaneously
- the most important unfinished project in crypto (in the eyes of many ogs at least)
- the last opportunity with an infinite tam (non-sovereign store of value)
- the most compelling price action setup (breakout after decade-long distribution)
confluence like this only comes once every few years
Happy retweeted
when i talk about dynastic things, i think of generational time periods
you can 躺平/quiet quit, check yourself out of society and genuinely enjoy the rest of life
but at the expense of handicapping your next gen, instead of giving them unfair advantages to destroy their peers
Happy retweeted
People forget Bitcoin ran from $800 to $20k in 2017.
This was during the dumpster fire of block wars, ETFs getting denied, shady exchanges, forever FUD on Tether.
Crypto liquidity was tiny then, not the trillion dollar oceans you see today.
Communism was not on the raise, Gov debt was not imploding globally, AIs had not constructed a panopticon.
Encrypted Bitcoin the last 1000x
Happy retweeted
The Tesla that Jessica just bought cost $48k and has a range of about 350 miles. The one she bought in 2015 cost $135k, adjusted for inflation, and had a range of 270 miles. People who think they dislike capitalism don't know how good they have it.
Happy retweeted
The mechanisms seem straightforward to me (and doesn’t require any speculation, just mild extrapolation.)
We’ve been getting almost daily AI breakthroughs in medicine. At its simplest I interpret the cause as “directed brute force.” If we can simulate enough protein folding and enzyme interactions we find helpful new ones that let us treat the body like a machine, intervening more and more as we gradually gain deep understanding.
Cancer treatments and cures are popping up all over the place.
“Cure all disease” and “cure death” are very strong statements. I don’t think anyone is arguing that AI will save your life if you jump into the sun. But if we look at mortality causes, AI is on its way towards eliminating ~90% of them pretty quickly (heart disease and automobile fatalities as examples.) We’re close to growing most organs in vats, have promising treatments for dementia/alzheimers/parkinsons. The only major cause of death that I’m aware of that remains mysterious is telomere shortening, but we know of many animals that have “solved” this, won’t take that long to understand how and apply it to humans.
Will AI in practice cure disease and death for all people? Of course not for political/social reasons. Same way people starve next to full supermarkets or die of drug abuse.
We’ve already gotten unbelievable medical breakthroughs in the last decade, but for political/bureaucratic reasons, your GP likely doesn’t know about many of them (like custom gene therapy for cancer), and there’s no distribution for a lot of this stuff.
Happy retweeted
the future is going to be amazing
don’t let the doomers psyop you into thinking otherwise
MODERNA, MERCK MRNA CANCER VACCINE SUCCEEDS IN PHASE 3
$MRNA and $MRK say their personalized mRNA cancer vaccine, intismeran, combined with Keytruda slowed melanoma recurrence and spread in a late-stage trial.
The study was stopped at its first interim analysis after positive results, though detailed efficacy data have not yet been released.
If the benefit holds up, this would mark the first randomized Phase 3 validation of a personalized neoantigen cancer vaccine and could significantly broaden the use of mRNA technology beyond traditional vaccines.
Happy retweeted
The outlook for medical breakthroughs that we are likely to see in our lifetimes, especially with AI and quantum, have to make you so bullish on life and humans broadly. Just try not to get hit by a bus in the next 10-15 years.
An experimental cancer vaccine kept melanoma from coming back. It’s the first successful late-stage mRNA study validating decades of custom-treatment research on.wsj.com/4qy7Ule
Happy retweeted
There was a brief moment when ethereum:0xa0df17b5ac76ababa36e1450e2cbcd18a620c845 looked like it would have 1000eth ($1.9m) in buybacks for August and traded at $4m marketcap.
Happy retweeted
Yeah terrible outcome for the industry generally, as is this practice of crying in the casino and supporting Burwick Law in suing anyone and everyone the moment your token does down.
Burwick are classic ambulance chasers, which is a term with originates from the practice of low-end and unethical lawyers chasing after ambulances to get to people that get hurt in car accidents to try and get them to sue someone.
E.g. you get in a car accident that isn't particularly anyone's fault and there are no serious injuries. Ambulance chaser lawyer shows up and convinces you to sue another driver, the city council, the manufacturer... anyone and everyone that might have money. They teach you to manufacture injuries (whiplash etc) and offer to do this for free, as long as they get a 50%+ cut of any payouts.
Their entire business model is not for you to win, but to make it costly enough to the random third parties that they pay money to go away.
This is exactly what Burwick does in the crypto industry. They wait till people lose money on something, anything, then try and sue anyone and everyone involved for the most outlandish and headline grabbing claims possible.
Following the memecoin mania, they have tried suing (from recollection), Pump, Meteora, Solana Fdn, Solana Labs, Jito alongside a whole bunch of projects such as Eliza.
The most telling fact is that Burwick never actually goes to trial on any of these claims. They simply drag it out for as long as possible and try to make it a big deal in the media on the (correct) assumption that it would hurt the defendants because (a) their reputations would be affected, (b) they have to spend millions of dollars on lawyers and (c) significant management time will be spent dealing with the lawsuits. The thesis is that if they inflict enough pain on you, you will pay them an amount of money to fuck off which they can frame as a win, and pay themselves.
In many ways, Burwick is a for-profit version of Gary Genslers SEC, where they threaten to kill projects with lawsuits, regardless of whether those lawsuits have any legal basis.
This quoted post is unavailable.
