Some Emails to the World
subj
Think chess is just a board game?
Think again. Scientists say the number of possible chess games is greater than the number of atoms in the entire universe — and that’s not an exaggeration. Known as the Shannon Number, this mind-melting figure (around 10¹²⁰) means that every move you make could open a new universe of possibilities — literally!
That’s what makes chess so endlessly fascinating. Even after centuries of play, no two games have ever been the same. From ancient kings to modern AIs, everyone’s still trying to master this infinite puzzle. It’s not just a game of strategy — it’s a glimpse into the boundless complexity of human thought.
That’s what makes chess so endlessly fascinating. Even after centuries of play, no two games have ever been the same. From ancient kings to modern AIs, everyone’s still trying to master this infinite puzzle. It’s not just a game of strategy — it’s a glimpse into the boundless complexity of human thought.
subj
Points on the board
make that "63 AND COUNTING" winners of The Nobel Prize - say the stickers on our football helmets for our game against UNC tomorrow...
Cal just won FOUR more Nobel Prizes - Go Bears!
Big brains in these helmets too 🧠🏆
Cal just won FOUR more Nobel Prizes - Go Bears!
Big brains in these helmets too 🧠🏆
subj
vibe coding as contral-signal of originality
So the inventor of the term "vibe coding" basically admits that unless your programming task is some bog-standard simple thing, "AI" will probably not make you faster or better (_if_ you take your work and responsibility seriously and actually check what those bots generate) but the opposite. Quoting Mr. Karpathy:
"I tried to use claude/codex agents a few times but they just didn't work well enough at all and net unhelpful, possibly the repo is too far off the data distribution."
So _if_ your AI makes you faster you are probably doing something by hand that an already existing library or framework could mostly do for you anyways with high code quality.
"I tried to use claude/codex agents a few times but they just didn't work well enough at all and net unhelpful, possibly the repo is too far off the data distribution."
So _if_ your AI makes you faster you are probably doing something by hand that an already existing library or framework could mostly do for you anyways with high code quality.
subj
11th century method for calculating Earth's radius!
Al-Biruni calculated the Earth's radius to be approximately 6,339.9 km (or 6,336 km in some sources), which is an incredibly close estimate for his time. He used a method involving trigonometry, the height of a mountain, and measuring the angle of the horizon's dip from the mountaintop. This calculation was highly accurate, with a margin of error of less than 1% compared to modern measurements.
Al-Biruni's method:
Concept: Al-Biruni understood that the Earth's curvature causes the horizon to appear lower from a mountaintop than from sea level.
Height measurement: He determined the height of a mountain by measuring the angle of elevation from two different points on flat ground.
Horizon dip: Using an astrolabe, he measured the angle of dip (or depression) of the horizon below the horizontal line from the mountaintop.
Calculation: He applied the law of sines to the large triangle formed by the Earth's center, the mountaintop, and the horizon to calculate the Earth's radius.
Result: His calculation for the radius was about 6,339.9 km, and his estimate for the circumference was within 0.3% of today's accepted value.
Al-Biruni's method:
Concept: Al-Biruni understood that the Earth's curvature causes the horizon to appear lower from a mountaintop than from sea level.
Height measurement: He determined the height of a mountain by measuring the angle of elevation from two different points on flat ground.
Horizon dip: Using an astrolabe, he measured the angle of dip (or depression) of the horizon below the horizontal line from the mountaintop.
Calculation: He applied the law of sines to the large triangle formed by the Earth's center, the mountaintop, and the horizon to calculate the Earth's radius.
Result: His calculation for the radius was about 6,339.9 km, and his estimate for the circumference was within 0.3% of today's accepted value.
subj
Story Vision: Duarte's superb advice!
After creating hundreds of thousands of presentations, Nancy Duarte discovered a framework in 2010 that changed her life.
She mapped it over Martin Luther King's "I Have a Dream" speech and Steve Jobs introducing the iPhone.
Both aligned perfectly.
She cried in her office - the pattern she'd been desperate to find was real.
See, most founder pitches fail the same way. You stack all the customer pain points at the start, then demo your product at the end.
By the time you reach your solution, people have already decided if they're interested. They tuned out at slide 8.
Duarte's Sparkline does the opposite. You alternate between “what is” and “what could be” throughout the entire pitch. Pain, solution. Pain, solution.
The pattern works because contrast commands attention and open loops create psychological discomfort. The brain needs recurring tension to stay engaged:
- MLK toggled between injustice now and "I have a dream" repeatedly.
- Jobs contrasted clunky smartphone limitations with iPhone capabilities throughout the 80-minute presentation.
- JFK alternated between the US’s space limitations and “we choose to go to the Moon in this decade.”
Each toggle made staying in the current state unbearable.
The execution:
1. Make your customer the hero by using their exact words
Interview five target customers or investors before you build slides. When they describe frustrations, use their language verbatim. This proves you understand their reality before pitching your solution.
2. Paint “what could be” with sensory detail
Not better accommodations. Instead: a family arrives in Paris, their Airbnb host left fresh croissants and a handwritten neighborhood guide on the kitchen table. They feel like locals, not tourists. Concrete outcomes stick. Abstract benefits are forgotten.
3. Alternative problem/solution throughout - never batch
Pain 1, solution 1, pain 2, solution 2, pain 3, solution 3. Never group all problems then all features. Batching lets investors and customers mentally check out before you finish.
4. End with an immediate next step (24-48 hours)
For investors: “By Friday, confirm the partner meeting date and three references you want to call.” For customers: “By tomorrow, send three use cases and I'll record a custom demo by Wednesday.” Make the decision immediate and concrete.
Watch for these signals mid-pitch:
You're losing them when investors lean back, check phones, or pivot to questions about your burn rate and competition.
You're winning when customers interrupt to describe their specific use case, ask about implementation timeline, or want to loop in their team immediately.
When every startup in your category has similar features, the pitch that creates unbearable tension wins the round, the sale, and the talent.
Act
She mapped it over Martin Luther King's "I Have a Dream" speech and Steve Jobs introducing the iPhone.
Both aligned perfectly.
She cried in her office - the pattern she'd been desperate to find was real.
See, most founder pitches fail the same way. You stack all the customer pain points at the start, then demo your product at the end.
By the time you reach your solution, people have already decided if they're interested. They tuned out at slide 8.
Duarte's Sparkline does the opposite. You alternate between “what is” and “what could be” throughout the entire pitch. Pain, solution. Pain, solution.
The pattern works because contrast commands attention and open loops create psychological discomfort. The brain needs recurring tension to stay engaged:
- MLK toggled between injustice now and "I have a dream" repeatedly.
- Jobs contrasted clunky smartphone limitations with iPhone capabilities throughout the 80-minute presentation.
- JFK alternated between the US’s space limitations and “we choose to go to the Moon in this decade.”
Each toggle made staying in the current state unbearable.
The execution:
1. Make your customer the hero by using their exact words
Interview five target customers or investors before you build slides. When they describe frustrations, use their language verbatim. This proves you understand their reality before pitching your solution.
2. Paint “what could be” with sensory detail
Not better accommodations. Instead: a family arrives in Paris, their Airbnb host left fresh croissants and a handwritten neighborhood guide on the kitchen table. They feel like locals, not tourists. Concrete outcomes stick. Abstract benefits are forgotten.
3. Alternative problem/solution throughout - never batch
Pain 1, solution 1, pain 2, solution 2, pain 3, solution 3. Never group all problems then all features. Batching lets investors and customers mentally check out before you finish.
4. End with an immediate next step (24-48 hours)
For investors: “By Friday, confirm the partner meeting date and three references you want to call.” For customers: “By tomorrow, send three use cases and I'll record a custom demo by Wednesday.” Make the decision immediate and concrete.
Watch for these signals mid-pitch:
You're losing them when investors lean back, check phones, or pivot to questions about your burn rate and competition.
You're winning when customers interrupt to describe their specific use case, ask about implementation timeline, or want to loop in their team immediately.
When every startup in your category has similar features, the pitch that creates unbearable tension wins the round, the sale, and the talent.
Act
express, not stress