debbiedowner

Does this scale to dining placemat size?

Tsla is 1.4T market cap, so it's almost like *ELON-stock is going to double in 1 day. It will go from 4% to 8% of qqq in 1 day.

It'll happen a week or a month after IPO date though? It took fb/meta 1 year and then it entered as 1% qqq. TSLA entered 3 years after IPO so probably a small percentage.

Tsla is 2% vti (2T AUM). QQQ is 400B AUM. So add those two and you get $56B of purchasing. This seems like the amount they want to raise via IPO in total in the news, so the banks who do the IPO can sell it all guaranteed.

But people will want to buy it before it gets into the passive funds... So... Post inclusion market cap will be higher than we expect?

How different is this to something you can buy like: https://www.amazon.com/ihreesy-Movement-Mechanism-Silent-Rep... ?

Checkout the movie "The Lost Leonardo" for a glimpse into the weird world of art attribution. tl;dw is that there are big financial incentives to attributing one way or another.

Why do you say rent control reduces the incentive to build more housing?

To me it seems the opposite: Rent control means supply goes down, so available building & land prices go up. These prices going up means an opportunity for builders who are good businessmen because they are going to make a margin on their investment, the bigger the investment the bigger upside.

Another intuition is with rent control it's hard to extract new value from an old building, so that also incentivizes tearing it down and squeezing more units into the land.

In SF, rent control exists on all buildings built before 1979. It appears to me that people who prioritize new builds pay a huge premium for them. I think this particular rule also incentivizes tearing pre 1979 structures down, vs the no rent control newer buildings can continue to have growth in the value extracted from them.

Assuming 200 acres is 0.57x0.57 miles. If you fly at 15 mph you can do 15 lines in 30 mins. The lines would be 200 feet apart, and you'd move 22 feet per second. 15MP images would be 10 pixels per foot and motion blur with 240fps would be 1.1 inches.

Anything wrong with that?

Better camera on the drone to fly faster and higher.

Has anyone tried it on a 3090?

Which books?

An LLM named Duet has been in Google docs for 17 months now! https://workspace.google.com/blog/product-announcements/duet...

I've been using it for about a year.

Could you make an onboarding guide for points?

My impression is if you fly for work, you get a lot of employer sponsored points, so it's interesting.

But if I fly 5-10 trips a year personally, why would I try points when I can get 3-5% cash back on my various cards?

Click on "available" and "employment" in "Waiting for AI Apocalypse / Available for Employment"

On the author's resume is a link to a L10 salary at google under "available for employment" and a distinguished engineer at Amazon. So author is in the top 0.01% (or even higher really) of salary expectations for the SWE ladder as I understand.

The link to the wivi research is to a 2013 article, so that took a while!

I had a HW assignment in 2017 to do this, specifically we were given raw wifi signals and told to determine how many people were in the room. We did it by detecting breath rate of each person with the MUSIC spectral algorithm. For that challenge there ended up being 6 people in the room, all with slightly different breath rate (attenuating the signal as their chest expanded).

Why 90%? Why don't you go on the cloud for training at least if it's not healthcare data?

Is it because you don't need to buy many gpus to do your workload?

We do something similar in house where I work. Is it hard to onboard new customers? Since they make a special container for you they basically adopt your build system, which may be hard for them. Does this go beyond mutation testing?

We add support by project, and the prototypical project we started with had 1M test reverse dependencies, a quarter of that was eligible test targets that we could recommend (based on language written in). This is probably the biggest project that we would ever find to support in the monorepo.

Some are UI tests, but we don't recommend those, because we found they don't catch breakages as often so we don't support the language they're written in. The tests we recommend are often integration type tests in that they call very higher order functions and often many of them.

This has been my job for over 2 years now!

We do it on a symbol level after statically analyzing each change, and everything in the monorepo daily. Our remedy to high risk changes is to run more tests, client tests not unit tests. Sometimes there are 100k client tests to pick, so we rank them and run a small subset.

It is a hard problem. One interesting observation is that there is a culprit symbol or two in the culprit change, but its connectivity is very similar to non culprits in the same change.

Another observation is that the transitively modified callgraph after a change is pretty big, a depth of 50 is not unusual. It is hard to get many useful signals out of it beyond amount of overlap in transitively affected symbols between change and test.

We found file level and build target level to be too coarse, but AST symbols are working.

Used in foiling boats in America's cup and similar races.

Doesn't Google cloud have 10s of thousands of employees? Isn't that too big to sell?

This is a shortcut/workaround to transforming the private docs to a prompt:answer dataset and fine tuning right?

What would be the difference in user experience or information retrieval performance between the two?

My impression is it saves work on the dataset transformation and compute for fine tuning, so it must be less performant. Is there a reason to prefer the strategy here other than ease of setup?

Just don't grade HW and have handwritten quizzes and tests in class? (Yes, the grad students will suffer when grading... unless you pay them and then they will be lining up)

What are some examples of the obsolete tools? I can imagine a manual hand drill...

What's your list of classic papers for language models for code, static analysis, or code summarization?

What about deep learning approaches like this one: https://ai.googleblog.com/2022/04/pathways-language-model-pa... ?

Or GitHub copilot?

I participated in a "ruby fight club" like this a few years ago hosted by grouper: https://github.com/Grouper/battleship

You can get a tutor? TAs are pretty money hungry so almost any of them will do it for 25 an hour.

I had a Prof say "no matter how I clown around up here, you're going to have to teach yourself these things".

My perspective is that classes provide the motivation, then each student does what they do with it.

I don't think departments are financially incentivized to create happy customers, and I do still think this is a good thing after my own multi-year "sufferings."

Mountain: inverse of a circle.

Molehill: reciprocal of all points in a circle on the complex plane.

Where is the Kickstarter?

When I look at his constantly increasing income I just think I would rather spend the money on a super good easy life and other people knowing there'll always be more.

I feel like living within your means makes sense both ways. Don't obsess materially and go into debt. But also don't obsess about the future enough to sacrifice the best possible present for yourself.