perhaps one of crypto's greatest disservices to humanity was making it culturally acceptable for builders to play shortsighted and negative sum games as a result, we've blown our lead and allowed for big banks and tradfi to come in, cannibalize the system, and create a new one not much different from the old

we will look back on bonk v. pump like we look back on sushi v. uniswap it is written in the stars

This is why, paradoxically, I’m almost sympathetic to people who cannot wrap their minds around the “revenue meta.” Not because these people are correct that it’s just a “meta”, they are very wrong and always have been, but because the first 10 years of crypto managed to completely nerdsnipe and indoctrinate so many in this industry into just writing open source code and assuming you’d capture value for ~reasons~ and the worst part was that it was *mainly true* for those 10 years, in other words, there was a positive feedback loop where these tokens *did* go up in value so people assumed they were on the right track, surely something wouldn’t go up in value if it was valueless. This was the wrong take away and now we reckon with the ramifications of this The fat protocol thesis brainfucked so many people into building infrastructure with no real path to a defensible moat or monetization and we are having to undo this insidious plague in real time. The whole “public goods” funding worked in the literal sense that it provided a lot of infra (even tho most of that infra is actually being siloed/ customized at this point), that part is probably true, but public goods *definitionally* do not make money, they are projects that would not be funded because because there is no financial incentive to do so. And a lot of god forsaken analogies were drawn to early OSS investment outside of crypto like red hat, mongodb, gitlab etc, but the strategy there to be very clear, was to write OSS and *then* sell it as SaaS like custom Linux or GitHub or database solutions / integrations / whatever. It wasn’t just hacking on a token and calling it a day which is what we did in crypto. And sure “building the infra first and then monetizing with the apps atop it later” is a reasonable strategy and has precedent in trad tech where acquiring users / building network effects before focusing on revenue later has worked well but this isn’t the same. Everyone was forced to hawk tokens that have a dubious value prop and people are realizing this now. The point though is we should be kind to, or at least understand, people who were deep in the sauce between 2016 to ~2023 who bought into this drivel, it was powerful and convincing to many and it made people feel like they were doing something different to make the world a better place and I at least understand why the idea of a project *actually* needing to make money is sort of jarring if you were brought up in this church

“nothing ever happens” prevails once again

this is getting out of hand

To all the VCs who read these tweets this all you need to invest in the next 10 years to 1,000,000,000x your capital. 1. Photonics - not for compute (not there yet), but anything in networking & IP. The IP is crucial here, the companies who win in photonics are not going to be selling chips they will be selling IP & Licenses. Photonics is going to be in every data center in the world in 2 years, probably even going to be used for longer distance communication let's see how this plays out. 2. Small Modular Nuclear Reactors (SMR) - data centers are consuming so much energy that at this point they need to be next to the power plant. Fusion is cool but what you want to wait 40 years? SMR for data centers is going to happen maybe in like 2 years honestly maybe even sooner. The power budget makes sense & data centers practically have nuclear power plant security anyway. Also delivering power creates LOSS, just put the nuclear reactor in the 2u rack man. 3. Computational Bio/Chem - no not because of Parm but she is doing good work. But because computational techniques for biology are so hard so difficult that we have no choice but to do lab experiements. Custom hardware & models here makes sense the value of accurately modeling chemical reactions is HUGE TAM (Pharmaceuticals, Oil, Plastics, Cosmetics) if you crack this and you can do it its not like we don't have the math your cutting out so many middle men & pain staking experiments. 4. Data - the new oil isn't compute, compute is going to be so cheap we don't care but data. GPT Wrappers would be multi billion dollar companies if they had data moats. You want to build some ai tool for like tire companies? You need tire data. You want to build an ai tool for airlines? You need airline data. Want to do it for restaurants? You need data. Compute is going to become a commodity we will be able to spin up b800 instances and crunch through terabytes of data but finding good data curating it and making it quality. THAT IS A MOAT. 5. Anything that makes semiconductor design faster & cheaper - not LLM Verilog I am talking about the hard stuff. I am talking about IP for silicon interposers, chiplets, but also just throwing compute at the problem. Many people are using AI for P&R but nope you misunderstand the problem. This is a classical optimization problem we need to use dusty old algos and bang them out. A coprocessor just for this is worth billions of dollars man it takes days sometimes weeks to clear timing on ASICs don't be a retard make this faster.

world is getting more chaotic by the day love it

easiest litmus test for intelligence/courage is if someone is capable of more than "post-game" analysis it doesn't take a genius to explain why something worked but do you have the intellect and courage to argue why something *might* work, and risk sounding stupid?

Great to see Alliance founders representing crypto on @tbpn and helping bring it into the broader tech conversation. We’re at an inflection point, AI has become incredibly competitive, and the smartest founders will recognize this and capture the asymmetry that still exists in crypto.

an asset is fundamentally worthless unless people believe in it why does tesla command 100x earnings while legacy auto struggles at 2x? it’s not superior fundamentals, it’s the elon narrative if apple's ceo died tomorrow, the stock would collapse even while revenue remained unchanged. it's the narrative shift, not the business reality that determines its value at the end of the day, we're all investing in narratives and prices are the metrics we use to value them markets are essentially where narratives are bought and sold and blockchain gives us the technology to price increasingly abstract concepts like ideas, trends, or beliefs the future enables pricing everything - from prediction markets to cultural movements - and if you onboard enough people, you have the most accurate financial meter for determining what a narrative is worth