Pages

Sunday, April 02, 2023

Just Applied To Y Combinator

In 2010, Paul Graham and I were featured in the same BBC article.





How Y Combinator Started I don't think we've ever managed to remember our birthday on our birthday. ......... The VC fund was doing what now seems a comically familiar thing for a VC fund to do: taking a long time to make up their mind. ......... As we turned onto Walker Street we decided to do it. I agreed to put $100k into the new fund and Jessica agreed to quit her job to work for it. Over the next couple days I recruited Robert and Trevor, who put in another $50k each. So YC started with $200k. ........... The company wasn't called Y Combinator yet. At first we called it Cambridge Seed. ........ Initially we only had part of the idea. We were going to do seed funding with standardized terms. Before YC, seed funding was very haphazard. You'd get that first $10k from your friend's rich uncle. The deal terms were often a disaster; often neither the investor nor the founders nor the lawyer knew what the documents should look like. Facebook's early history as a Florida LLC shows how random things could be in those days. ........ We started Viaweb with $10k we got from our friend Julian Weber, the husband of Idelle Weber, whose painting class I took as a grad student at Harvard. Julian knew about business, but you would not describe him as a suit. ............ In return for $10k, getting us set up as a company, teaching us what business was about, and remaining calm in times of crisis, Julian got 10% of Viaweb. I remember thinking once what a good deal Julian got. ............ we wanted to learn how to be angel investors, and a summer program for undergrads seemed the fastest way to do it. No one takes summer jobs that seriously. The opportunity cost for a bunch of undergrads to spend a summer working on startups was low enough that we wouldn't feel guilty encouraging them to do it. ............. The structure of the YC cycle is still almost identical to what it was that first summer. ............ We never expected to make any money from that first batch. We thought of the money we were investing as a combination of an educational expense and a charitable donation. But the founders in the first batch turned out to be surprisingly good. And great people too. We're still friends with a lot of them today. ............ It's hard for people to realize now how inconsequential YC seemed at the time. .......... Jessica and I invented a term, "the Y Combinator effect," to describe the moment when the realization hit someone that YC was not totally lame. When people came to YC to speak at the dinners that first summer, they came in the spirit of someone coming to address a Boy Scout troop. By the time they left the building they were all saying some variant of "Wow, these companies might actually succeed." .......... it took a while for reputation to catch up with reality ....... That's one of the reasons we especially like funding ideas that might be dismissed as "toys" — because YC itself was dismissed as one initially. ........ The density of startup people in the Bay Area was so much greater than in Boston, and the weather was so nice. ........ Plus I didn't want someone else to copy us and describe it as the Y Combinator of Silicon Valley. I wanted YC to be the Y Combinator of Silicon Valley. So doing the winter batch in California seemed like one of those rare cases where the self-indulgent choice and the ambitious one were the same........ we didn't have time to get a building in Berkeley. We didn't have time to get our own building anywhere. The only way to get enough space in time was to convince Trevor to let us take over part of his (as it then seemed) giant building in Mountain View. .......

The first dinner in California, we had to warn all the founders not to touch the walls, because the paint was still wet.



at No comments:
Labels: , ,

Anand Sanwal: The Vision Man Of CB Insights

I have been a long time subscriber to Anand's newsletter, easily over a decade. I feel like his particular expertise is a vastly under-monetized resource. He data crunches his way to amazing insights into many tech sectors today and his projections into the future are also insightful.
at No comments:
Labels: ,

Two Martians Walk Into A Bar

Two Martians walk into a bar, and Elon Musk turns to the other Martian and says, "I heard Earth is flat." The other Martian looks at him incredulously and says, "Elon, you're an alien! You've seen the curvature of the universe with your own eyes!" Elon just shrugs and says, "Eh, I guess I just haven't been paying attention." The bartender overhears this exchange and chimes in, "Well, I guess this bar is the only place flat-earthers are welcome."



at No comments:
Labels: , , ,

Saturday, April 01, 2023

1: Lex Fridman

at No comments:
Labels:

1: Big Ben



GPT-4 Is a Reasoning Engine Reason is only as good as the information we give it ......... there are at least two important components to thinking: reasoning and knowledge. Knowledge without reasoning is inert—you can’t do anything with it. But reasoning without knowledge can turn into compelling, confident fabrication. .......... Even though our AI models were trained by reading the whole internet, that training mostly enhances their reasoning abilities not how much they know. And so, the performance of today’s AI models is constrained by their lack of knowledge. .......... GPT models are actually reasoning engines not knowledge databases. .......... if you want to make an investment that indexes the success of companies building in AI as a whole, one smart move would be to invest in a vector database provider, or a basket of them. (Alternatives might be to invest in OpenAI, or a basket of large cap software companies like Microsoft and Google that build AI, or chipmakers like NVIDIA that build the GPUs that AIs run on.) ......... AI’s advancement is a raindance that calls forth capital from Patagonia vest wearing angels. ....... People have been saying that data is the new oil for a long time. ........ We tend to underappreciate the significance of the input—what information we feed it to produce those results. Its answers are largely dependent on the information we make available to it for analysis. It’s only as powerful as its starting point. .



How to avoid war over Taiwan A superpower conflict would shake the world ......... Europe is witnessing its bloodiest cross-border war since 1945, but Asia risks something even worse: conflict between America and China over Taiwan. Tensions are high, as American forces pivot to a new doctrine known as “distributed lethality” designed to blunt Chinese missile attacks. Last week dozens of Chinese jets breached Taiwan’s “air defence identification zone”. This week China’s foreign minister condemned what he called America’s strategy of “all-round containment and suppression, a zero-sum game of life and death”. ........... Is it willing to risk a direct war with another nuclear power to defend Taiwan, something it has not been prepared to do for Ukraine? And by competing with China militarily in Asia, could it provoke the very war it is trying to prevent? ........ China could use “grey-zone” tactics that are coercive, but not quite acts of war, to blockade the self-governing island and sap its economy and morale. Or it could launch pre-emptive missile strikes on American bases in Guam and Japan, clearing the way for an amphibious assault. Since Taiwan could resist an attack on its own only for days or weeks, any conflict could escalate quickly into a superpower confrontation. .......... a war over Taiwan could involve a new generation of arms, such as hypersonic missiles and anti-satellite weapons, causing untold destruction and provoking unpredictable retaliation ........... Taiwan is the world’s essential supplier of advanced semiconductors. America, China and Japan, the three largest economies, and among the most interconnected, would deploy sanctions, crippling global trade. America would urge Europe and its other friends to impose an embargo on China. .......... War is no longer a remote possibility .......

President Xi Jinping has told the People’s Liberation Army to be ready for an invasion by 2027, says the CIA.

.......... The military balance no longer so clearly favours America as it did in the 1990s. And public opinion has shifted in Taiwan, not least because of how China has snuffed out freedoms in Hong Kong. Only 7% of Taiwanese favour reunification. .......... Both sides are shoring up their positions and trying to signal their resolve, with destabilising consequences. ......... the mysterious severing of undersea internet cables to remote Taiwanese islands. .......... Top American and Chinese defence officials have not spoken since November. During the recent spy-balloon incident, a “hotline” failed when China did not pick up. ......... The island is admirably liberal and democratic, and proof that such values are not alien to Chinese culture. .......... A botched invasion, however, would cost Mr Xi and the Communist Party dearly. ........ The goal should not be to solve the Taiwan question, but to defer it. .......... The first 15 years of the American-Soviet cold war featured a terrifying mixture of brinkmanship and near-catastrophic mistakes, until the Cuban missile crisis prompted a revival of diplomacy...... the two rival systems must find a way to live together less dangerously.



Google Bard is about to get supercharged the current metaphorical Lebron James of large language models: GPT-4........ Training a large language model on outputs from someone else's large language model is a little like photocopying a map and calling yourself a cartographer — not a good look.



Trump was indicted. We collected the best tweets about it.
at No comments:
Labels:
Subscribe to: Posts (Atom)