Had a really interesting conversation with @matt_levine about the intersection of technology and finance, coming to an Internets near you in the not too distant future.
I definitely learned something; hope everyone will enjoy it.
Specific interesting thing from about the first minute: I asked Matt what a convertible bond was for the audience and he gave the true technical answer (a bond which can be surrendered for stock at some agreed upon terms) and why it was interesting to companies.
That reason: you should only convert the bond above some share price, which effectively means you have a call option on buying the stock. Call options are more valuable if volatility is high. This is well understood.
Therefore, a borrower running a business with high variance (like a tech company with unproven model or a biotech firm) can get a loan which they’d probably not qualify for under traditional underwriting standards (“Where’s your cash flow?!”) by bundling implicit option premium.
That’s kind of beautiful: the very fact that your business is risky de-risks the offering enough such that you can access the money necessary to take your risky shot.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
Somewhat self-indulgent because I'm mentioned in the thread, but yeah, this is a concrete example of Google getting convincingly outcompeted in search, *which should surprise me* and no longer does.
I haven't yet replaced my core uses of Google Search with "just ask an LLM already, they have a non-zero chance of getting it right" but I think there are a lot of searches for a lot of people where that is already obviously the optimal play.
One very common genre for me is "tip of the tongue" searches, which for common knowledge questions I've already moved entirely to ChatGPT, but for "can you find a document I read in mid-2000s with the following properties" I still default to considered use of Google.
I think Twitter porn bots are something like a visible test of non-state capacity.
Similar to the old line about McDonalds restrooms’ cleanliness: easy to check, hard to fake, surprisingly difficult to solve at scale, solvable at scale given sufficient capacity.
And when I think of capacity degradation at Twitter under current management the people who expected it to failwhale and then go hard down permanently were clearly wrong and the people who expected no impact to capacity in firing X0% seem to also be wrong.
Guess what will very predictably happen if I click any button here.
Both an interesting visual and a strategy for bootstrapping an event, from Less Online.
a) Write a list of who would attend the event in an ideal world.
b) Show people that list. There, that’s the pitch.
c) Color green anyone who says yes.
d) Repeat, expanding list as necessary.
Solving the cold start problem is a thing you can just choose to be good at.
SBF said at least one useful and true thing, which I will paraphrase. Sometimes somebody needs to say “There is a fixed time and a fixed place where everything will be ready for the magic to happen, except you. Can I count on you being there?”
Some years ago I did a speech in Tokyo at a simultaneously interpeted Japanese/English. Simultaneous means attendees get an earpiece connected to an interpreter who is doing very-close-to real time interpretation.
Due to the wishes of the event organizer and the fact that most speakers could not speak Japanese, I was asked to do my presentation in English. Fine for me, saves a lot of effort in preparation.
Me to a company: Just following up here, when are you going to do the thing.
Company: We have done the thing.
Me: Your doing the thing would be visible to me, but it is not yet visible. Have you done the thing?
Company: Lol you might be bad at looking.
Me: Always possible.
Me: I have rechecked and cannot find evidence of you having done the thing. Strictly hypothetically, if you have not done the thing, there needs to be a resolution path. What would that look like.
Company: We did the thing.
Me: Hypothetically.
Company: Hypothetical, blah blah.
Me, in writing: Previous to our conversations and upon commercially reasonable expenditure of effort I cannot find evidence that you have in fact done a thing you must do. I request that you engage your resolution path.
Company: Oh funny story.
Note also that you don’t have to build a career somewhere you care about. This is a classic issue for young people. (Video games, many forms of saving the world, etc.)
You can build a career doing *something you can cause yourself to care about* where *most people could not.*
The world will forever need someone to really sweat a bunch of boring details buried under reams of boring reports and explained by people who do not attend the best parties. It will tend to compensate people who do that work pretty well, particularly if they care about that.