Pinned Tweet
"In this choiceless never-ending flow of life, there is an infinite array of choices. One alone brings happiness: to love what is." ~Dorothy Hunt
15
3
49
The best thing I did this week was read a full article without prompting it through AI, thinking deeply about it, and exercising my brain
Felt I haven’t done that in a long time
Felt I haven’t done that in a long time
1
0
3
195
Why Pearl will Win
––––––––––––––––––
01:14 BTC is dying
03:24 Breaking up with my highschool girlfriend
06:51 Why it can work
09:35 Pearl's power footprint
12:45 AI capex > asic capex
15:00 FP8
19:00 Negative overhead
22:04 Elite team
23:06 Hashrate = hinge profile height
28:01 AI nt my MTV
32:31 Mark Knopfler shreds on guitar
35:52 Useful work vs Duplex work
38:32 Cheap verification is required for money
41:20 Can Pearl close the loop?
45:17 Addressing fundraise FUD
51:21 Risks
1:06:27 Pearl wins by becoming multi-denominational religion
2
0
4
250
Morning Mist retweeted
Just bought a fat bag of $NU / @nubank here:
Insane -30% sell off due to the unexpected results of the election in Brazil right now, entire Brazil market is bleeding.
Generational entry in an hypergrowing fintech business with one of the sharpest founders @velez_david in the 39% YoY with quarterly earnings of 1.1 BILLION (!!).
139 million users. Sure, 85% of it are still in Brazil and hence the heavy reaction to the elections, but already massively growing their market share in Mexico and Colombia too, slowly but steadily becoming the leading neobank for Latam.
Based on a fair forward P/E, the mid-term target sits anywhere between $19-29 per share, currently trading at $11 pre-market.Fearful market that is overindexing on the second order effects of this election, while it's ignoring the actual numbers for one of the most exciting and hypergrowing neobanks.
As always, expecting this dust to settle pretty quickly and wouldn't be surprised if we see $NU bouncing 50-100% mid-term, with quarterly reports scheduled for Nov 12th.
Insane -30% sell off due to the unexpected results of the election in Brazil right now, entire Brazil market is bleeding.
Generational entry in an hypergrowing fintech business with one of the sharpest founders @velez_david in the 39% YoY with quarterly earnings of 1.1 BILLION (!!).
139 million users. Sure, 85% of it are still in Brazil and hence the heavy reaction to the elections, but already massively growing their market share in Mexico and Colombia too, slowly but steadily becoming the leading neobank for Latam.
Based on a fair forward P/E, the mid-term target sits anywhere between $19-29 per share, currently trading at $11 pre-market.Fearful market that is overindexing on the second order effects of this election, while it's ignoring the actual numbers for one of the most exciting and hypergrowing neobanks.
As always, expecting this dust to settle pretty quickly and wouldn't be surprised if we see $NU bouncing 50-100% mid-term, with quarterly reports scheduled for Nov 12th.
22
4
92
10,755
On-chain vesting and clawbacks are the most underrated founder filter. If a team won't accept auto-enforced investor protections, that's the diligence answer. You don't need another reference call.
solana:METAwkXcqyXKy1AtsSgJ8JiUHwGCafnZL38n3vYmeta
solana:METAwkXcqyXKy1AtsSgJ8JiUHwGCafnZL38n3vYmeta
0
0
0
79
You've got to adapt to a changing world - revisit your priors, and they did.
Smart IMO
Smart IMO
Scoop: a16z's flagship blockchain fund plans to drop "crypto" from its name and rebrand as a16z frontier.
a16z crypto was one of the first and largest mainstream blockchain funds, raising a monster $4.5 billion vehicle in 2022, and more recently a $2.2 billion fund earlier this witz also became one of the most vocal—and deep-pocketed—defenders of crypto in Washington, D.C.
Longtime general partner Ali Yahya is also leaving the fund.
a16z crypto has already been investing in non-blockchain startups like Kalshi and the AI compute market startup Ornn, and this move to expand beyond crypto echoes recent strategies by other top blockchain firms including Paradigm and Haun Ventures.
Still, general partner Chris Dixon has long been a public face for the Web3 movement, and the decision represents a symbolic shift for one of the most prominent crypto venture operations.
Full story here: theinformation.com/articles/andre…
0
0
0
86
Recently read Demian by Hermann Hesse, one of my favorite authors.
I loved this conversation between Sinclair (protagonist) and Demian:
Sinclair: “Can you really make another person think what you want him to?”
Demian: “No,” he said, “that can’t be done. That is to say, one n if the person acts that way. Neither can the other person think as he will, nor can I make him think what I want him to..."
“To take an example from the butterfly world, there is a certain species of moth, of which the female is much rarer than the male.
Suppose you have in your possession a female of this type of moth, naturalists have often made the experiment, then the male moths fly in the night to this female, they even make a flight of several hours’ duration! Think of it! For many miles around all the males are conscious of the whereabouts of the only female moth.
Moths must have a sense of smell, or something like it, which allows them to pick up and follow an almost imperceptible scent, like a good hound...
If an animal or a man concentrates his whole attention and his whole will-power on a certain thing then he attains it. That’s all. And it is just the same with what you have asked me. Observe a man sufficiently well, and you will know more about him than he does himself.”
Sinclair: “But what has that got to do with will-power?” I asked. “You said there was no such thing as free will. And then you said one only had to concentrate one’s will on something to be able to attain one’s ends. That doesn’t agree! If I am not master of my will, then I can’t direct it here or there as I wish.”
Demian: “A good question!” he said, laughing. “You should always ask questions, you must always doubt. But the explanation is very simple. If a moth for instance wants to concentrate his will-power on a star or something like that, he can’t do it. Only, he doesn’t try. He seeks only what has sense and value for him, satisfies his needs, he gets what he absolutely must have.
To be sure, I can imagine this or that, or make myself believe that I absolutely want to get to the North Pole or somewhere, but I can only carry that out and wish it strongly enough when the desire lies right in myself, when my whole being is really filled with it. As soon as that is the case, as soon as you try to carry out an inward command, then you succeed, then you can harness your will as you would a good nag."
I loved this conversation between Sinclair (protagonist) and Demian:
Sinclair: “Can you really make another person think what you want him to?”
Demian: “No,” he said, “that can’t be done. That is to say, one n if the person acts that way. Neither can the other person think as he will, nor can I make him think what I want him to..."
“To take an example from the butterfly world, there is a certain species of moth, of which the female is much rarer than the male.
Suppose you have in your possession a female of this type of moth, naturalists have often made the experiment, then the male moths fly in the night to this female, they even make a flight of several hours’ duration! Think of it! For many miles around all the males are conscious of the whereabouts of the only female moth.
Moths must have a sense of smell, or something like it, which allows them to pick up and follow an almost imperceptible scent, like a good hound...
If an animal or a man concentrates his whole attention and his whole will-power on a certain thing then he attains it. That’s all. And it is just the same with what you have asked me. Observe a man sufficiently well, and you will know more about him than he does himself.”
Sinclair: “But what has that got to do with will-power?” I asked. “You said there was no such thing as free will. And then you said one only had to concentrate one’s will on something to be able to attain one’s ends. That doesn’t agree! If I am not master of my will, then I can’t direct it here or there as I wish.”
Demian: “A good question!” he said, laughing. “You should always ask questions, you must always doubt. But the explanation is very simple. If a moth for instance wants to concentrate his will-power on a star or something like that, he can’t do it. Only, he doesn’t try. He seeks only what has sense and value for him, satisfies his needs, he gets what he absolutely must have.
To be sure, I can imagine this or that, or make myself believe that I absolutely want to get to the North Pole or somewhere, but I can only carry that out and wish it strongly enough when the desire lies right in myself, when my whole being is really filled with it. As soon as that is the case, as soon as you try to carry out an inward command, then you succeed, then you can harness your will as you would a good nag."
0
0
1
155
My main learning from studying RL
GRPO, the best method, is free and identical for everyone
Whoever has the best tasks, the most realistic environment and the most trustworthy check or judge for a valuable job trains the best agent for that job
That’s the moat
GRPO, the best method, is free and identical for everyone
Whoever has the best tasks, the most realistic environment and the most trustworthy check or judge for a valuable job trains the best agent for that job
That’s the moat
0
0
0
136
A founder told me they run daily tests for agent readability online, and are spending up to 50% of their time optimizing it to get "in the token flow"
Their product shows up in 4% of relevant agent queries. That number is the new SEO benchmark
Optimize it or die
Their product shows up in 4% of relevant agent queries. That number is the new SEO benchmark
Optimize it or die
0
0
0
101
Morning Mist retweeted
building Turnkey has given me a unique view into what actually works in consumer crypto.
we’ve worked with some of the biggest apps in the industry, from Axiom to Moonshot to Etherfi, and I’ve gotten to see firsthand what separates the apps that win from the ones that don’t.
here’s how to ACTUALLY win in consumer crypto:
1/ I cannot emphasize enough how important it is to survive the cycles.
crypto is one of the most volatile industries out there, and in the span of two years the market can go up, down, and sideways multiple times over.
a few strong weeks of trading volume can make new founders feel like the money will keep coming forever, especially right after a raise. it won’t.
you have to manage burn carefully, stay disciplined, and make sure the business can survive long enough to catch the next wave.
2/ you must have a defensible moat.
we’ve seen the same crypto concepts launch over and over again, and a lot of them end up failing for the same reason: there’s nothing defensible about them.
in crypto especially, you need a real moat. that can come from the product itself, a technical advantage or even distribution and brand.
Moonshot simplified the consumer buying experience.
Axiom offered lightning-fast trading for crypto power users.
Etherfi gave robust neobanking to a global audience.
you must figure out what your moat is and build around it for years.
3/ think long term.
one of the biggest mistakes founders make is optimizing for the next few months instead of the next few years.
it’s easy to get distracted by token launches, incentive campaigns, short term volume or whatever narrative is hot that quarter.
those things can accelerate growth, but they can’t be the entire business.
the best companies focus on building products people will still want to use when the incentives disappear.
4/ distribution matters just as much as the product.
you can build the most technically advanced product with the best UI in all of consumer crypto, but none of it matters if nobody knows it exists.
the best way is to build viral loops directly into the product. many of these have been perfected by non-crypto products: invite incentives, Venmo-style “Your friend sent you $5” flows, and social network effects
launchpads are a perfect example of this. every new token launch naturally creates something for users to share, promote and rally attention around.
social trading products work the same way. trades, positions and leaderboards become content that gets pushed back into the feed, bringing more users into the product.
to sum it up... consumer crypto is hard, but the playbook is simple:
survive long enough, build something defensible, think in years, and make sure distribution is part of the product from day one.
we’ve worked with some of the biggest apps in the industry, from Axiom to Moonshot to Etherfi, and I’ve gotten to see firsthand what separates the apps that win from the ones that don’t.
here’s how to ACTUALLY win in consumer crypto:
1/ I cannot emphasize enough how important it is to survive the cycles.
crypto is one of the most volatile industries out there, and in the span of two years the market can go up, down, and sideways multiple times over.
a few strong weeks of trading volume can make new founders feel like the money will keep coming forever, especially right after a raise. it won’t.
you have to manage burn carefully, stay disciplined, and make sure the business can survive long enough to catch the next wave.
2/ you must have a defensible moat.
we’ve seen the same crypto concepts launch over and over again, and a lot of them end up failing for the same reason: there’s nothing defensible about them.
in crypto especially, you need a real moat. that can come from the product itself, a technical advantage or even distribution and brand.
Moonshot simplified the consumer buying experience.
Axiom offered lightning-fast trading for crypto power users.
Etherfi gave robust neobanking to a global audience.
you must figure out what your moat is and build around it for years.
3/ think long term.
one of the biggest mistakes founders make is optimizing for the next few months instead of the next few years.
it’s easy to get distracted by token launches, incentive campaigns, short term volume or whatever narrative is hot that quarter.
those things can accelerate growth, but they can’t be the entire business.
the best companies focus on building products people will still want to use when the incentives disappear.
4/ distribution matters just as much as the product.
you can build the most technically advanced product with the best UI in all of consumer crypto, but none of it matters if nobody knows it exists.
the best way is to build viral loops directly into the product. many of these have been perfected by non-crypto products: invite incentives, Venmo-style “Your friend sent you $5” flows, and social network effects
launchpads are a perfect example of this. every new token launch naturally creates something for users to share, promote and rally attention around.
social trading products work the same way. trades, positions and leaderboards become content that gets pushed back into the feed, bringing more users into the product.
to sum it up... consumer crypto is hard, but the playbook is simple:
survive long enough, build something defensible, think in years, and make sure distribution is part of the product from day one.
11
2
84
4,706
LPs love our crypto returns.
But strip out one token launch and MOIC drops to 1.1.
The real risk is not crypto: It is drifting into AI at inflated vals to chase the narrative.
But strip out one token launch and MOIC drops to 1.1.
The real risk is not crypto: It is drifting into AI at inflated vals to chase the narrative.
0
0
1
135
A Latam crypto infra company bootstrapped from low single digit millions to near eight figures.
Over two dozen clients across a dozen countries.
Never took VC. Simple website, crypto OG founder, zero SF norms.
That's the real compounding nobody is tracking.
Meeting them friday, lmk if you want in
Over two dozen clients across a dozen countries.
Never took VC. Simple website, crypto OG founder, zero SF norms.
That's the real compounding nobody is tracking.
Meeting them friday, lmk if you want in
1
0
3
396
Sometimes founders are blinded by their relationships with coworkers
CEO hired sales guy, a friend, sales guy hired 2 guys, didn't get a single sale in the last 4mo
CEO is hesitant about firing, investors sometimes are there to reassure these decisions over some drinks
CEO hired sales guy, a friend, sales guy hired 2 guys, didn't get a single sale in the last 4mo
CEO is hesitant about firing, investors sometimes are there to reassure these decisions over some drinks
0
0
1
302
Neobanks in LATAM have PMF but can't grow ARPU.
Startups going after credit engines have a huge opp here.
Add credit infra as an SDK, take a cut of interest, and your distribution problem is solved.
Startups going after credit engines have a huge opp here.
Add credit infra as an SDK, take a cut of interest, and your distribution problem is solved.
0
0
1
267
Pearl turning AI matrix multiplication into proof-of-work is the most interesting protocol design i've seen this year
The main problem, 99% of it it's not for inference right now (see chart)
In Q4 FIP8 will allow inference of frontier models, and that *supposedly* will be the network
Miners earn from securing the network AND producing compute
The main problem, 99% of it it's not for inference right now (see chart)
In Q4 FIP8 will allow inference of frontier models, and that *supposedly* will be the network
Miners earn from securing the network AND producing compute
3
0
12
1,595