Liquids are heavy. A large chunk of the retail cost in bottled drinks of all types is transport and distribution.

The same is true of liquid detergents and soaps and anything else you buy that is mostly water - even fresh produce.

Where do you store high quality original images in case future edits or recompression with better codecs are needed? Generation loss is a thing.

I view the high-quality originals as “source” and resized+optimized images as “compiled” binaries. You generally want source in Git, not your compiled binaries.

As always, it really depends on what the source is. Often images are created with some software like Photoshop, would you commit the psd file to git? If you're a photographer, would you commit 20mb+ raw image files? Might make sense for a few images, but git is just not the right solution for binary data in general. Every modification has to duplicate the entire file. This makes working with the repo unpleasant very quickly.

In general, I recommend people back up binary files to cloud storage, like S3, and only commit optimized, deployment ready assets to git. There's also GitLFS, but it's clucky to use.

I like this take. I tend to agree.

> you can just enable either scx_lavd or scx_bpfland from the kernel settings

So Linux is still nowhere near an option for non technical users.

It just depends on one distro to default on scx_bpfland.

For technical users, it's already the best option.

You’ve never had a concussion, gone under anesthesia, or gotten older? Memory based passwords are not durable; I personally forgot my Google account password after a surgery.

I grew up in a rural township 50 miles from a major city in the 1980s. We were never isolated and there were in fact a diverse set of peers my own age with interests and heritage all across the spectrum. Yes there were a few racists or religious zealots but 99% of the folks got along just fine.

My own lasting impression is that this is the “American experience” that is not dead nor impossible to recreate in 2026. We just all need to learn to be decent Americans again.

> I know other people like reading LLM output.

I haven’t met any of these people; I’m sure some may exist but does anyone actually “like” rather than “tolerate” LLM writing? Anybody have a link to a decent study or survey on this area?

> AWS designs and implements their foundational services holistically.

I’d say they implement their services circularly. The outage-inducing circular dependency between Dynamo and Route53 is not a “holistic” design.

SendGrid and their competitors are already the very definition of “sender pays” for email. “Sender pays” is how they make money. This isn’t a problem of monetary incentives.

The problem is that companies get their SendGrid credentials compromised via password re-use or phishing.

I mean the carrier pays the recipient , so Twilio and sendgrid bear some cost

They understood just fine. But because that cost passes through to the sendgrid customer, it wouldn't motivate sendgrid to stop enabling spam.

currently the costs are too low to affect policy. that's my point. and the recipients are making extremely high margins on ads, so they don't have much reason to push back, either.

For any reasonable email fee, sendgrid can continue passing it on to the customers and not care.

If you make the fee super high, then many email workflows completely break and sendgrid goes out of business.

I don't think there's a number where it does what you want and incentivizes sendgrid to be careful.

(And you might say to seek a middle ground, but I don't think there is one. My guess is that "too low for sendgrid to care much more about a couple percent of mail from hacked accounts" and "too high for sendgrid to still attract customers" probably overlap.)

spam volume is 10000x-1e6x higher rate, so even small fees would impact them much higher than legit senders.

These are the accounts of legit senders being coopted to send very targeted spam. I don't think you can distinguish it by volume, because the volume needed to make these schemes work is just a fraction of the basically-legitimate volume these services process.

The real bulk bulk spam is a different issue entirely.

It never was FIPS-approved and likely will never be. The wireguard protocol used by Tailscale uses ChaCha20 for encryption which is not FIPS approved.

Interesting. What is the FIPS version of wireguard?

There are some forks that are not compatible with regular wireguard, for example from wolfssl. Or just classic mTLS.

> What is the FIPS version of wireguard?

IPsec or TLS-based overlays which use AES encryption and NIST-approved ECC curves or (gasp) RSA for key exchange and authentication. They generally suck in comparison with wireguard, which is a clean-sheet modern cryptographic protocol.