kenhwang

So far it's just Ohio that had gerrymandering backfire, turning +2R into +2D.

Texas, Florida, and NC are still too close to call to see whether gerrymandering netted any new seats.

The good ones still make time to knock out some lifelong backlog resident tickets to sharpen their axe on.

The bad ones get promoted for spending the time instead playing politics.

Matches my experience in adtech, Q4 was always at least half of our yearly revenue. Q1 usually came in at less than 1/4 of Q4.

That's my understanding as well. Seems like they're improving a $20 part of a billion dollar process, and a part that was already heavily using state of the art AI/simulation at that. The computers already spit out more potential candidates than what can be reasonably tested in vitro.

The new AI powered Alexa beta basically can't do egg timer, music, or light switch duties as reliably as the old Alexa, which basically makes it useless.

You'd think with all the data collected by these tech companies, they'd understand the common usage scenarios of their products. But instead we have AI Alexa that can't egg timer, or Gemini Assistant that can't navigate or make phone calls.

Rust has Rhai and Roto too if you need better than Python level of developer considerations.

I'm hoping one day people look at Python the same way they look at Perl and PHP.

There has been so much programming language innovation in the past decade that I'm surprised someone hasn't built a better Python. I'd like to see something like Rhai or Roto take off which has a natural upgrade path to Rust.

Maybe it's because interpreted languages just aren't in vogue anymore and probably won't be for a long time if humans aren't even coding anymore.

They're not competing against each other, they're competing against free/local/open weight models that they're also lobbying very hard to regulate/ban.

Ruby caught a second wind when they started putting effort into their JIT, which in our testing made it quite a bit faster than every interpreted language except JS in real world performance.

But then Python got a second wind in performance with better rewrite in Rust integration, but it looks like with the latest DHH keynote, Ruby might also go down that path.

My opinion is that I/O and concurrency ease of use and performance continues to be Ruby's weak point vs other interpreted languages. Which also happens to be Go's very few strong points.

Most banks let you schedule recurring transfers.

It's also marketing that their AI is hilariously terrible because AI Overview is the biggest posterchild for AI slop and often used as an example of how bad Google is at AI. Gemini proper isn't that bad, just the ultra cost cutting version on search is.

The ones in LA just get stuck behind things because LA loves its park anywhere you want anarchy.

But also the Waymos use alleys, driveways, and private parking lots that they're really not supposed to.

Always a bit embarrassing when the Waymo decides to use a space that's not meant for public use for pickup/dropoffs. One time, mine waited for me in/blocking a fast food drive-thru lane.

Pretty sure most of the post-exposure (PEP) cost is from the immunoglobulin, which is also why antivenom is often so expensive because it is also often immunoglobulin based. A synthetic alternative to immunoglobulin should bring the price down and make it easier to store/transport.

Pre-exposure (PrEP) for rabies (a vaccine) is more expensive than it should be due to lack of economics of scale and the typical American medical pricing inflation. That's not unique to rabies, it's just how vaccines are allowed to be priced in the US.

My understanding is poorer rabies prone countries do not even attempt PEP for rabies due to the cost, they really rely on broad PrEP immunity.

PEP is always going to be more expensive than PrEP, because transplanting a vaccinated animal's immune response proteins is a much harder thing to do than getting your body to make it's own.

They get stuck in LA as it is which is pretty chill for a major American city all things considered.

They're just really cautious and do need an extra 10s to think sometimes, and there's a lot of driving scenarios that don't allow that much time to decide to go or not. The biggest culprit seems to be unprotected intersections with the potential for human/bike crossing, which includes driveways which they insist on using for dropoffs (even when they're obviously not public driveways), which they then take a very long time to leave.

The grass too, everyone knows bikes ride on grass.

My previous employer even suggested code reviews are for human code only, AI-only code can go out without review as long as CI passes since the AI can be prompted to review its own code before committing.

Arguably hybrids rely on software much more than EVs do since it has to all the motor/battery management EVs do, but also coordination with the engine and transmission which EVs do not have, and Toyota has been the leader in that area for almost 3 decades now.

Unless you mean software as in fart apps on the touch screen, then yeah, Toyota lags in that department.

I'll say track and test yourself, since it depends highly on skin tone and weather.

That study also claims 12-20ng/mL 25-hydroxyvitamin D is sufficient. My doctor says the post-COVID studies suggest 75ng/mL is closer to optimal, and <30ng/mL had been considered deficient for about a decade before COVID.

Personally I had no problem hitting 20ng/mL through sunlight alone and I'm pretty fair skinned but live somewhere sunny. But I haven't been able to get anywhere near 75ng/mL even when I've made an effort to increase vitamin D through supplements and exposure to the sun.

The studies I've seen from HCl claim it's more bioavailable too, so you only need 1/3 the dosage compared to honohydrate. That makes the cost pretty comparable.

But otherwise, I'd agree that monohydrate has significantly more research behind it and I'd default to that if it's tolerable.

I haven't seen the 20g loading dose recommendation on anything recently either. Everything I've seen just recommends 5g/daily.

My regular supplements that I've noticed made any difference is: creatine, vitamin D, and fiber. With vitamin C and protein being situational as needed.

I don't think dosing affects much, as long as you have excess creatine, the extra water requirements are determined by your muscle mass.

20g is a loading dose (typically for a week), then it's 3-5g daily for maintenance. But I've seen studies that 5g daily catches up to the 20g loading dose after about a month.

My understanding is it's a bit of a bind and release process that moves the water from outside of the muscle cell to inside of it until the cell is saturated or the creatine is consumed.

So the amount of creatine changes the rate as which water is loaded into the cell, but the total amount of water needed is calculated based on your muscle mass. The figures I've seen is 2L extra water (spread evenly, not all at once) for a normally recommended 5g daily creatine intake during loading (which typically takes a month at 5g/day dosage).

My doctor added vitamin D test to his patients recurring blood work since the start of COVID, and found out just about all his patients was deficient or near it.

If you don't have a primarily fish based diet or spend hours under direct overhead sunlight every day without a shirt on, you're probably deficient as well.

I get it too if I don't drink a lot more water. So did many people on my sports team.

Creatine binds to water (making less water available to your muscles) and allows your muscles to work harder (which makes them need more water). So if you're not drinking a lot more water when you're taking creatine, you will get cramps from either dehydration and/or muscle overuse.

But once you're past the loading phase for creatine and getting sufficient hydration, it's supposed to lessen cramps because the water bound to creatine also becomes available for use again as the creatine is used by the body.

I just wish Graphene had better support for non-Google hardware.

More that my motherboard came with a Broadcom WiFi chip instead of MediaTek which everyone recommends for Linux, and the shipped Linux drivers can only use 1 of the 4 antenna and won't connect to any 5ghz networks.

It "works" with 2.4 with 1 antenna, but the speed is then capped at 150Mbps instead 600Mbps/1300Mbps the hardware is capable of.

So it's like buying an AWD car, but realizing only one wheel gets power. Still technically drivable, but hardly working well. Maybe most people wouldn't even notice at all, but I figured I was technical enough to get in the weeds and fix it. Not all cars end up being Toyotas, sometimes you get a Jaguar.

Nailed it. My Linux machine was built from the retired parts from my main desktop, so it very much is stuff that served me well for several years under Windows.

But the advice is still very much to research the hardware or buy a prebuilt with vendor support.

Compared to a decade ago, more stuff just works; especially noticeable with graphics drivers. But when things hiccup, which they still do, it still requires quite a lot of tech savviness to solve.

One of my machines is always Linux desktop; I've been using it to see if it's ready enough for my main desktop to switch over and so far my verdict has been it's not.

Last weekend I tried updating to Linux 7.2, and WiFi drivers failed to compile, so it's been fun.

For me, it's gaming, drivers, and still too much Windows-only software.

Wine/Proton has gotten really good, but still requires tinkering, and still sometimes has really bad performance.

WiFi/BT drivers continue to be really terrible, there's connectivity but speed and reliability sometimes aren't there. Peripherals and motherboard chipset functionality still mostly won't work.

Desktop environments are still clunky and rough around the edges. Maybe I don't get along with KDE/Gnome and just need to shop around for another DE.