Brian

@krismicinski I used to do the exact same in my compiler class when Macs were x86, but stopped when they were ARM. Which flags are required to get it to run on MacOS?

@krismicinski macOS x86? Are you running on ancient hardware or do you have a trick for getting Rosetta to work?

@AndrewCMyers @mgill25 Agree. I cover parsing because I feel students are likely to write some sort of parser in the future. But I don’t like the material that much.

@bdemsky @mgill25 it's true that you can spend a lot of time on parsing if you want to. The existence of decent parser generators these days means it's not so worth it.

@AndrewCMyers @mgill25 To be fair, I have seen compiler class schedules that spend a lot of time on parsing.

@mgill25 And let me just add that I actually do care about parsing. It's enabling to know how to write clean parsing code, and the techniques have intellectual depth. But there is so much else to cover...

@mgill25 Sounds like you are reading the wrong textbooks.

@roydanroy If you want to take phone calls on them, the Sony headphones have bad microphones.

@roydanroy AirPod Max. Tried a bunch of them.

I'm considering noise cancelling headphones for an open office situation. Does anyone have thoughts on Sony WH-CH720N Noise Cancelling Wireless Headphones versus JBL Tune 770NC Happy to hear about better choices (in similar price range).

@JAldrichPL One question is what are the critical products that will not be on store shelves? It seems TP is made in America, so hopefully I don't need to fill my garage with that. :)

The Trump Tariffs are actually Covid to our economy. Hang on, it's going to be rough ride.

@TaliaRinger Sorry to hear. Did she start day care? That was the worse for us for illnesses.

It just must be so hard to experience these things and not know what they are. Like how can she even know this isn't just forever now?

Baby is sick and so miserable, I feel so bad for her. I had this bad cold before her and it sucked. I'm just getting over it and she's at the worst part now

@satnam6502 Xilinx?? What is the software culture like there? As a software guy who has had to use their tools, they manage to make it extremely difficulty to version control FPGA projects. It is like crazy difficult and any approach gets broken with a new version.

This is the Silicon Valley I moved to in 1998, leaving my academic position at the University of Glasgow to join a silicon chip company called Xilinx (now AMD). It was a very different time, a very different vibe. I miss it.

@GuthmannR @JustinWolfers Which index is the Europe line tracking?

@JustinWolfers The thing is that up to 2013, US and EU stocks were similarly performing; from 2013-2024, a big divergence emerged. Maybe now we will have correction to convergence:

This chart seems to suggest that something is going wrong in the United States, specifically. (ft.com/content/7568e7… )

@TaliaRinger Mine never did. Daycares tend to be flexible, but some often want you to call if you are late. I think primarily because they want to make sure parents don't forget their baby in the car while at work.

Do all babies shift to the magical mythical 7-7 sleep schedule, or are some babies just night owls? I can't imagine putting my 4.5 month old to sleep at 7, and if I do she just treats it like a nap anyways. Sweet spot is like 930 PM right now. Wondering where 7 PM comes from lol

@farmerrf Do you realize that an overhead rate of x means that x/(1+x) of the total budget goes to overhead?

There is a big shakeup coming at top tier universities that rely on Federal funding through NIH and NSF grants. Recently announced changes have cut the overhead rate on these awards to 15%. Before the change, UCLA charged 57% Princeton, 67% and Harvard charged a whopping 69%. Like most academics, I spent a good part of my career seeking research funding. Unlike an experimental physicist or a chemist, I do not need a laboratory to carry out my work. Back in the day, I simply needed a room, a blackboard and a piece of chalk. A typical NSF award would pay to support my graduate students, some travel funding to conferences and a buyout of my time from teaching. Today, substitute a new computer for the chalk — but at today’s prices that’s a rounding error on a typical NSF award. A typical award for a top tier theoretical economist might cost the NSF $500,000 over three years. At an overhead rate of 69%, $345,000 of that award would go to the university to provide the blackboard, the room and the chalk and the remaining $155,000 would pay a teaching buyout and graduate student support at less than $50,000 a year. Rethinking the way the Universities are funded is a timely move.

@TaliaRinger @AndrewCMyers @gimpel_israel I don’t like overhead rates either. But my students like having an office with heating and cooling and electricity. And the government has bunches of rules that have to be checked when spending money. And all of this costs something.

@AndrewCMyers @gimpel_israel But I think the model of universities mostly getting money via absurd overhead costs on grants that faculty apply for is in place as a form of tax evasion

Deep cuts to the National Science Foundation only make sense if you are an enemy of science or of the US. Most basic research in the US is funded by the NSF. The whole NSF budget is less than 0.2% of the federal budget, and most of it goes directly to scientists.

@roydanroy I have definitely seen discussions of this on online local Llama forums such as reddit. I believe one thing they mention is that for the really large model sizes, it runs very slow.

Has anyone out there compared the GPU/memory specs for the new Macbook Pro M4 Pro/Max computers to the requirements to run inference in the more interesting open source LLM models, especially those aimed at reasoning? What I'm curious about is whether there's a compelling case to spring for a Macbook Pro M4 or one of the Pro/Max model from the perspective of being able to do "inference" in one of the more interesting models on the laptop itself, without having to deal with an external machine/cloud, etc. Would really appreciate any insights people might be able to offer.

@roydanroy Do you have a tail event?

I believe the tail risks associated with raising children are too large to make it worth any individual's while. If you disagree, are you 1) confusing societal and individual welfare? or 2) have no experience with a tail event?

@AndrewNoymer Is this a consensus opinion? Like what % of the field believes this?

Epidemiologists, collectively, have not been vocal enough about the origins of covid.

@JAldrichPL I could see a potential issue with publication. Supposed researchers started to feel pressure to experiment on themselves to get a paper? Getting this sort of experiment approved to be done on someone else would involve a lot of paperwork.

There's no legitimate ethical issue here. If the paper is scientifically worthy it should be published! A shame the article had to wait so long to see the light of day due to squeamish editors.

@TaliaRinger Change the time zone and set the appropriate time.

To feed the baby I need an alarm set for the first 130AM, not the second one after the clocks turn back, and I think this is literally impossible on my phone? lmao

@krismicinski I also wonder whether eventual AGI solutions might use more money in electricity than a programmer’s salary.

@krismicinski That may very well be true. But I don’t think LLMs will get us to that point.