Like Bloop, we’re also focused on modernization, but our approach extends beyond code to include the people behind these systems and capturing the institutional knowledge they hold.
We're using Maitai's structured output in prod (Benchify, YC S24) and it's awesome. OpenAI interface for all the models. Super consistent. And they've fixed bugs around escaping characters that OpenAI didn't fix yet.
Have you tried finding allies in the faculty on this issue? I strongly reccomend you find a handful who share your concerns and leverage them to get this madness resolved.
Sharing desks is hard, I'm curious if they are trying to use COVID WFH policy to justify over-accepting students, hoping the schedules just magically work out. Fun fact, they wont. No amount of metrics will stop two people wanting the same desk at the same time, if allowed. Or am I mistaken? Why else would monitoring desk usage be on the table (excuse the pun)? Surely they aren't trying to use these figures in evaluating a students work, right?
I think it's pretty sad that Northeastern spent so much money building a fancy new science center, which is mostly unusable covered space, and now there aren't enough desks.
We have some excellent faculty allies. The admin is interested in doing the same thing to faculty (making them share space) and so faculty is 100% on our side. Hence we had a public letter with 250+ signatures including many faculty.
I agree they likely over-accepted and spent money poorly but I can't say for certain, as I'm not in admin and thus not privy to these details.
GATE, but with M140 ATACMS, drones, and more competency. Also, no fan service because everyone involved with that in the real show dies in the initial shock and awe bombing.
I'm a PhD student in CS in the USA. At my school in Boston, we are paid a ~40K USD/yr stipend. My friends in industry make a _minimum_ of 120k/yr, and some make considerably more than that (think 200k+), in junior / "entry-level" positions.
When I complete the PhD, if I go to industry in the USA, my income will probably be similar to that of my friends who will have been in industry the entire time (and gotten steady wage increases throughout).
It's important to note also cost of living. 40k/yr might sound like a lot, but in Boston, rent is >1k/month even with roommates, we don't have dental care, our health insurance is imperfect, groceries are expensive, etc. etc. Meanwhile in Tucson Arizona or Bloomington Indiana the stipend is something like 22-28K/yr, as cost of living is lower.
Generally speaking it's reasonable to say that completing a PhD in computer science is not a financial investment, but rather, something I am doing because I want to do it. I am very unlikely to literally "profit" (compared to, if I had gone straight to industry instead).
A friend of mine got a phd and he told me the main benefit was that he had lots more opportunities. He was always called back for job interviews and had lots more positions available. His wife, a nurse, basically supported him while he was getting the degree.
This is still an opportunity cost scenario. Would your friend get called back for job interviews if he had just spent that time in industry and had ~3yr experience on their resume (My experience is yes).
Yes, but in the time it takes to complete a PhD, the junior engineer can usually get promoted at least once. So you both end up in the same place, but the guy who went straight to industry was making 3x more than the PhD in the meantime.
Don't forget that a PhD makes it more difficult to get some kinds of jobs. An intermediate or low level developer job may pay more, but if you have a PhD the interviewers are going to question why you are applying for a low level job and be worried how long you will stay. So in some respects a PhD reduces your employability (though you can always lie and say you never got the PhD, though it is kind of hard to hide on a resume without leaving a time gap, which makes you seem even less employable.)
Why would you even consider using an unsound logic?
I’m a 3rd year PhD student in formal verification and for the life of me I can’t imagine why anyone would choose to use an unsound logic. I must be missing something very obvious. Is it, IDK, “sound as long as you don’t reason about some very specific kind of formula which can be easily avoided”, or something like that?
The usual justification, within PL Design community is that, similar to Russell 's Paradox, it's said that the anomaly caused by the unsound logic does not come up in practical situations. For example, some (and definitely most if not all mainstream) languages admit `Type: Type` i.e. `Type` itself is a type within the formalism. Of course, this is unsound and one common work-around is to use universe polymorphism such as `Type n: Type (level-suc n)`. But this comes with its own design issues as now you may need to pass level to each type (depends on the language, situation etc). So, some decide, since (as per this reasoning) absurd situations implied by "Type: Type" don't come up in practical situations, it's better have a simpler and more understandable yet unsound Type Theory. (But of course, an inconsistent theory can prove any claim, which is a huge problem in and of itself)
I have to note that I personally do not agree with this reasoning, yet afaik this is one of the justifications used to introduce rules to TT of a PL that is known to be unsound.
Realistically, it all depends on what you want to do with the type system. If you want to have a general purpose programming language that is not restricted by logical technicalities not well-understood by programmers, maybe it's better to admit "Type: Type" or similar. If you want to have a dependently typed programming language where any runtime program can potentially be a compile-time program, you may need more guarantees (e.g. you may need to know certain functions will halt).
Because, as that quoted sentence points out, the logic itself is simpler and the creators think that makes it easier to use for teaching.
Taken to an extreme, if all you want to show off is how proofs are constructed, then whether or not your logic is actually sound doesn't matter. Since the goal of the tool isn't correct proofs, the proofs don't need to have any soundness properties.
When you create a new logic with certain applications in mind, it might be unsound at some intermediate steps, like you can have bugs in a program. But yeah, I don't think there is much use for an unsound logic per se.
What kinds? What’s the best process to do program synthesis? The more I write programs by hand the more I think it should probably be a first-class design consideration for new programming languages
I recently made a map (using Generic Mapping Tools) of one of the Aleutian Islands in Alaska, for one of my dad’s grad students to use in a paper. That would constitute a good example of where you couldn’t just tell my # of built structures (since at certain zoom levels, there are none!)
Sure, but are you likely to have users want to A/B test traffic light configurations there? My suggestion was very specific to the application in question.
You’re right. There are two types of 0: the internal syntax denoting digit shift, like in 507, and the number representing nothing, as in 0. So in our number system, o isn’t just syntax, it’s also a number.
