En 2005, Jeff Bezos dio una charla de 50 minutos en Stanford.
20 años después, muchas de sus ideas siguen sonando adelantadas a su tiempo.
Algunas que vale la pena recordar:
1. Cuando te acostumbras a un problema, dejas de intentar resolverlo.
Bezos lo llamó “indefensión aprendida”.
2. No elijas entre dos buenas opciones.
¿Servicio al cliente o precios bajos?
¿Por qué no ambos?
3. Haz más experimentos.
No te obsesiones con que cada idea parezca brillante.
Aumenta el número de intentos.
4. Obsesiónate con el cliente.
Los competidores te dicen qué están haciendo.
Los clientes te dicen qué deberías construir.
5. Si todos creen que tu idea es absurda, no la descartes demasiado rápido.
A veces, las ideas más extrañas solo necesitan más tiempo.
En ese momento, Prime llevaba apenas una semana.
Costaba $79 al año.
Y Bezos planteó una pregunta que hoy parece obvia:
¿Por qué no convertir la entrega en 2 días en un coste fijo?
En 2005, no era nada obvio.
Hoy, casi parece sentido común.
Las mejores ideas de negocio suelen parecer absurdas justo antes de volverse obvias.
Jeff Bezos gave a 50-minute talk at Stanford in 2005.
20 years later, most of the advice still sounds ahead of its time.
A few ideas worth stealing:
1. Problems become invisible when you get used to them.
Bezos called it ���learned helplessness.”
2. Stop choosing between two good options.
Customer service or low prices?
Why not both?
3. Run more experiments.
Don't obsess over making every idea look brilliant.
Increase the number of shots you take.
4. Obsess over customers.
Competitors can tell you what they're doing.
Customers can tell you what you should build.
5. If everyone thinks your idea is stupid, don't automatically kill it.
Sometimes the weirdest ideas just need more time.
Prime was only a week old when Bezos gave this talk.
$79 a year.
And his question was basically:
Why shouldn't 2-day shipping just be a fixed cost?
That question sounds obvious now.
It wasn't in 2005.
That's the point.
The best business ideas often look ridiculous right before they become obvious.
@sotraxyz Fair game: stack size is almost the whole story. Biased game: it isn’t. 10 units at 52% vs a 10x opponent is ~55% survival, not 9% with a rounding-error bump. The lesson that survives the formula is still bet size, not “edge doesn’t matter.”
Most people spend years learning what to say.
Almost nobody spends an hour learning how to make people listen.
Patrick Winston taught at MIT for decades.
His “How to Speak” lecture breaks down the things great speakers do differently:
→ Start with a promise
→ Use the board, not just slides
→ Make ideas visual
→ Never hide your hands
→ Know when to stop
→ Leave people with something worth remembering
No fancy tricks. No motivational fluff.
Just decades of teaching distilled into one hour.
Watch this once.
You’ll never look at a lecture, presentation, or speaker the same way again.
@MadCrash_X Once you split a rate into time + inflation + risk, “6.9%” stops feeling like weather and starts feeling like a quote you can actually check.
The biggest AI opportunity might have nothing to do with chatbots.
Watch what happened with AlphaFold.
A problem scientists had worked on for ~50 years became something AI could tackle at extraordinary speed.
Now Demis Hassabis is thinking much bigger:
→ games became the training ground
→ AlphaGo proved AI could search beyond human intuition
→ AlphaFold took that idea into biology
→ the next step: using AI to accelerate discovery itself
Drug discovery. Materials. Climate. Biology.
And eventually, perhaps AGI.
The fascinating part isn't the hype.
It's the possibility that AI doesn't just make existing work faster — it changes how fast humanity can discover new things.
This Cambridge lecture is worth an hour.