masterjack

Please explain in your own words why you believe this is not the case.

It wasn’t 2023: Last post 11 months ago, last comment 8 months ago, which is a typical level of lurking

I thought they were projecting 10B and said a few months ago they have already grown from a 1B to 4B run rate

Remarkably there’s really just one way to tie them together, you can always manipulate the knot to move between the different variants

13/183=0.071 so how can it be 20pct for this round??

I may agree if it was a 20% dilution round, but not if they are increasing from 3% to 7% dilution. Being so massively oversubscribed is a bullish sign, bad companies would be struggling to fill out their round.

It will typically be the same 40 GB model loaded in, but called with many different inputs simultaneously

Yes, you can `print axioms` to make sure no axioms were added, make sure it compiles with no warnings or errors. There’s also a SafeVerify utility that checks more thoroughly and catches some tricks that RL systems have found

It’s possible that in the Google deal you had to agree to sell back the shares (at a low value like par or original strike price) and the 1% refers to either those proceeds or the size of the Google employment package. If you didn’t agree then you would be left holding your shares of a company that is now gutted.

In this case P6 was unusually hard and P3 was unusually easy https://sugaku.net/content/imo-2025-problems/

Exactly. And presumably had a more sophisticated harness around the model, longer reasoning chains, best of N, self judging, etc

Have you found https://sugaku.net/ useful? It’s focused on math research

Which is fine for steady companies, but perpetually drags down any rapidly growing company

There’s also the search and browsing on https://sugaku.net, it’s more focused on math but does also have all of the arxiv on it

I also like Cline since it being open source means that while I’m using it I can see the prompts and tools and thus learn how to build better agents.

As a long-time Schwab user, one thing I was shocked by is that the login flow silently truncates the password to eight characters. I found this since I tend to have complex suffixes I rotate around, and one day I was able to login with the wrong suffix and even with no suffix at all. This must be due to some legacy process only allowing eight characters.

It'll be that price multiplied by how many SaaS tools you use, each of which will have some premium for their Enterprise plan supporting SSO, and typically they won't be transparent about what the price is or how to sign up for it, so you'll then have to go through the steps of negotiating an enterprise plan. Spending a few minutes to figure out their API sounds more fun

Are you really suggesting they open up 20 different bank accounts? And not only pay the monthly account fees but also pay a bookkeeper to upload statements for all 20 banks into the accounting system? And have appropriate controls on all money movements? And float money between the accounts each time payroll or some large expense is run? And follow the financial results of all of those banks in order to find signs of weakness?

Id rather be in a world where businesses don’t have to spend so much time playing games with their bank accounts and just trust that their money is safe, which is why the fdic needs to guarantee the deposits.

Shooting up and destroying key infrastructure for a town leaving many without power sounds a lot more like domestic terrorism than free speech, especially if it is in response to a small number of adults having an event you disagree with.

Blocking selling is a good way to get sued, so brokerages are very hesitant to do it. If the price drops the trader can claim they wanted to sell and are owed the difference. Blocking buying doesn’t have this issue since you can’t sue for theoretical gains.

True, but that’s the interesting thing about these kinds of algorithms, they align the pricing approaches so the entities act more like one than they did before, effectively decreasing competition.

It’s true, but it’s not like other markets where someone can choose not to buy and additional sellers can come in, to make sure that prices are in some ways tied to costs. Here, everybody has to live somewhere, so by default you have a “naked short” position on housing. Efficient price discovery here is performing a short squeeze and finding the point of “maximum you can afford to pay without going broke”, which is only possible to reach with price collusion.

Anyone know if this acquisition is a win for Plaid, or are they are just panicking before bank API middleware becomes commoditized (e.g. http://paysli.com/ offering free authentication)

Fragmentation is definitely high. Although if you can seamlessly fall back to other services, a handful of bank integrations can get you ~80% of the customers.

I haven't worked with Open Banking, how is it? In the US, for bank transfers, NACHA files are fairly easy to work with, as long as you can access and verify bank account numbers (ideally get balance and identity information too). I think all the APIs support this, eg: https://developer.wellsfargo.com/apis/payments/account-valid...

It's hard to tell how quickly things will change and improve. Banks should have an incentive to work on this, since it will give them more control of their data and logins, potentially lowering liabilities and hack potential.

What is Plaid's long-term competitive advantage? It made sense when scraping was the only solution (albeit an extremely dangerous one) but now banks are launching their own APIs, which is allowing some competitors to come in with free bank account linking. A few simple features address most of the needs of a typical fintech, and I don't think anyone is happy with having to integrate Plaid's SDK nor pay their rates.

Ruining the good name of Ruinerwold

In addition to the sibling’s comment on explicitly passing inna comparator, note that std::map is an ordered dictionary and requires an ordering on your elements. It sounds like you may have preferred hash_map

Yes but photosynthesis doesn’t generate energy but the opposite: it takes in energy from the sun in order to reverse the reaction

Given that it was a data science interview is it possible they were expecting it to be a fast simple thing (some pandas, some sklearn, you’re done) and instead he did some long slow manual thing so they rejected him?

Hedge funds can also be a good testbed for technologies, while “do good” companies can use it as an excuse to be bad on the inside. I am fascinated with the potential of large scale crowdsourcing in data science, and finance seems the most natural domain for this, which is why I’m now at Numerai.

Regardless of intent, SIPC has around $3 Billion saved up to protect all brokerage accounts in the US. They literally cannot afford to insure Robinhood if they wanted.