Profile avatar
joshuaotti.bsky.social
Hi, My name is Joshua Cook Otti. I'm a PhD student at UT Austin studying theoretical computer science and the author for stemforest books. My website: https://www.cs.utexas.edu/~jacook7/ Stem Forest Books website: stemforestbooks.com
299 posts 118 followers 55 following
Regular Contributor
Active Commenter

Read Ryan's work putting time t in space sqrt(t). Embarrassed, I didn't think of it. It's just the Ian Mertz Tree Eval paper applied to Turing Machines. One needs to reduce generic computation to tree eval, but I've seen the techniques before in PCPs, so not a good excuse.

If Ryan's new paper is accurate, this may be one of the biggest advancements in the since the PCP theorem. people.csail.mit.edu/rrw/time-vs-...

I think PCPs could benefit a lot from formal verification. They are very technical, delicate and complicated. It seems like lot's of things aren't known about them simply because it requires a lot of engineering, and theory papers aren't a good way to verify engineering type challenges.

So I have to fly for a job interview in Berkeley soon. Suddenly, much more nervous about flying then I have ever been before.

Getting stuff prepared for STEM girl day this weekend. Hopefully I'll inspire a future mathematician with error correcting codes.

Got a random order for my book from someone in Florida. Interesting. Glad to make a sale.

On a recent paper I set out to make a simple PCP construction that simply optimizes a couple of the less optimized parameters. With one composition. For simplicity, I constructed two similar PCPs to compose. A reviewer suggested it would be better to construct it once. I regret listening.

Sometimes I try to think of ways to make PCPs from expander codes. No success as of yet.

I'm going to tell people with Teslas the same thing I tell people with gas cars. You don't need to rush to replace it, just don't buy another one. Selling your tesla doesn't lose Tesla (much) money and making a new car is bad for the environment.

One thing I did not expect after completing a PhD in computer science is that I would be less competitive on the CS market. I'd figure companies would at least interview me, but no.

The onion always has to be so over the top to be clear satire. But banning words like equality, bias, inclusion, and privilege in the NSA, an agency centered on security (with extensive cryptography related math)? Wait, that one was real?

I don't think we should rebuild after disasters if we know the area is likely to be hit again. Either move, or change the environment to lower the risk. Don't rebuild LA, half of Houston shouldn't be rebuilt, same with many coasts in Florida.

So I had a conversation with Claude the other day about errorncorrecting codes, and it did surprisingly well. It came across as someone who is knowledgeable about error correcting codes, just not my specific problem.

I'll probably get an industry job when I graduate, which is sad, but I've also been under financial stress lately, so it should be a relief. I will miss math. Industry probably will be software engineering because IDK who wants someone who studies error correcting codes or interactive proofs now.

At least I got a paper into STOC before I graduate and I should finish my CCC paper before the deadline.

I am looking at how much I need to write in the next month with dread. Hopefully most of it will be line by line corrections and then copy and pasting papers I've written into a thesis. But, I'm worried that it won't be that easy.

Comic Idea: The devil is in he'll because he's actually really bad at contracts and when he can't fulfill his end of the bargain, he is on the hook for whatever sins were committed on his behalf.

Of the Trump executive orders, pulling out of the "OECD’s Global Tax Deal" irritates me the most. It is not the worst, but it is the most hypocritical. It's an American first policy that encourages business in America and only hurts rich tax dodgers.

Many people don't realize the dodo was around 500 years ago. We killed it. It's not a prehistoric bird. It is a modern bird we hunted to extinction, then we blamed it for dying.

The way we unload airplanes is a great example of how being individually polite leads us to being collectively inefficient.

This is how prison should work for most people. They get sent to rehabilitate. If economic security is a factor in the crime (as is often the case) they get trained for a new job (that actually has demand). I still think they should pay more and there should be more opportunities like this.