randomsolutions

He is rejecting the framing of get in now before it's "too late". If it is so useful then we will be able to pick it up when it is more polished rather than learning to use some half polished turd that will be obsolete in 6 months.

The only reason I haven't switched, is when using uv as my environment manager, starting/connecting to the python kernel in notebooks within vscode takes forever.

My biggest issue is using uv envs in vscode under WSL. Starting up interactive sessions takes forever. Its just too slow, can't figure out what the deal is.

I've been at two small companies that had layoffs, and who was cut/retained was 100% based on what was best for the company (except some visa holders were retained). Which is what it should be.

That is not what the study found. The study found people are able to guess the given names of adults from four choices at slightly better than chance, ~30%. With 117 (or 116?) people guessing and 16 images.

I agree it seems like flimsy justification. But it is also likely harder to assess and communicate. Temperature they get a point prediction for the high and you can easily calculate the mean absolute error.

For precipitation you will be getting percent chance often with an interval, 10% chance of 0.1-0.25 inches with higher likely in thunderstorms. Also precipitation patterns tend to be much more irregular within small spatial extents. You can asses things like calibration and perhaps take a mean value for there intervals to get point errors. But all of this will make it harder to communicate actual performance.

spud guns are timeless

Can't wait for the "Great Thermostat War"

Hopefully the idea you are trying to convey has slightly more coherence than predicting individual words.

I like the ping pong of one day an article being posted where everyone asks, "when/why did everything become so complicated", and then the next day something like this is posted.

Time to shack up. Double your income and halve the demand.

Why would you compare starting salary to the median home price? Are you suggesting every home below the median is uninhabitable?

The median home price would suggest that the median household income is at least $160k, which seems pretty good.

Yea, I don't see the problem with someone rediscovering knowledge.

The question I would ask is, would geohot have been better to have gone to school to learn this and deferred all the work he has done?

Who is farther ahead, the people who already knew this from school, or the guy who is building and learning this as he goes?

To who? As opposed to?

My thought too, or just a P controller.

Porous concrete has been useless. It clogs up eventually and loss it's ability to transport water.

The groundwater disappearing is a problem if you build up ag/industry/development based on the groundwater availability and then lose that source of water with no economical alternative.

I keep seeing these semi-religious Bayesian posts pop up. I would like to add a more nuanced perspective by none other than the great Bayesian practitioner Michael I Jordan.

I use Bayesian methods often, but this a just religious. Bayesian methods are just that, tools, methods for approaching a problem.

There are no laws for applying probability to the real world. To think so puts too much faith in your models. Remember, all models are wrong. Applying probability to the real world requires a host of assumptions, regardless of the methods you use.

Frequentist and Bayesian methods have different goals, both have there place.

For a counterweight to the strong likelihood principle find discussions of Larry Wasserman: https://youtu.be/Z-YvWyM6dRQ?si=qwzRiaPbj9ruiUEv

And for a balanced discussion for why both are great see Michael Jordan: https://youtu.be/HUAE26lNDuE?si=cwg6wpRS1gXL6r1Y

Don't listen to them this is awesome.

I think it is awesome. Definitely glad every page isn't like this, but it is cool, and goes with the vibe of the blog.

Unfortunately leaving the tab and returning seems to make it crash. But I'll take this over another striver blog.

You're describing bandwagoning. Cargo cult is misattribution of cause, not follow the leader.

Sucks for the 10% waiting 478 hours.

Search is one of the least interesting applications of LLMs. Most of the complaints about search are self imposed, not technical problems. Why are we still talking about a 20 year old problem that is basically solved?

The big takeaway from these articles that are coming out should be that baseline estimates are really difficult to do. For any carbon project that is deferring harvest you need to be able to accurately predict what would happen if an area was not included in the project, that is, predict a future that will not occur.

Project developers may use one method to do this, rating agencies may use another method, but at the end of the day we can never actually know. There are so many unknown physical, social, and economic factors that go into whether or not some area will be harvested. This is why we see such huge ranges in estimates of impact.