Notices by John Ripley (jripley@mastodon.social)

Excited to hear that we’re going to rename all the Washington and Jefferson streets everywhere.

What is this nonsense narrative that “US opposition to vaccines is growing”? No it isn’t. It’s a tiny minority of people, almost entirely those infected with MAHA. It’s also due to an active disinformation campaign. It’s not “covid backlash” either—that was the *same* disinformation campaign by largely the same characters.

Fucking media doing everyone a disservice by not just saying the actual obvious truth.

https://www.reuters.com/business/healthcare-pharmaceuticals/doctors-bear-burden-medical-freedom-fuels-worst-us-measles-outbreak-30-years-2026-02-13/

@inthehands I'm uneasy about such a rushed, disorganized general strike, because if it ends up being a dud, it might get taken as a signal that support is waning. I would even say that it's unhelpful, because it takes away energy which could have been used for something more organized. But on the other hand, "all PR is good PR". I'm hoping for a good outcome.

RE: https://hachyderm.io/@inthehands/115957156730683593

Reminder that all the DHS, ICE, CBP, and TSA, have only existed ~25 years, as an authoritarian power grab post-9/11 by the Republican Party, and they could easily *stop* existing. Their function wasn't necessary before, and isn't now.

@ryanc Spent a few minutes puzzling over why it's so bad, and it looks like that era of Atom was just before Intel greatly improved latency/throughput. Also I think it has only one pipe doing all the above, mostly stalled. I got into this thread because one of the last things I did at my prior employer was hyper-optimize AES-GCM for their cores, and they run bytes-per-cycle, not cycles-per-byte. Literally 10x faster.

@ryanc I’m suspicious maybe Intel checked the “AES-NI” box of ISA support in Atoms, while the underlying implementation is whatever is enough to pass validation only.

@ryanc That chip apparently has AES-NI, so AES-GCM should, if code is using it, be expected in the ballpark of ~2 cycles per byte, i.e ~1GB/sec.

Thought of the day, and I wish there were a way to get this to legislators:

Come the next Big One earthquake, all of San Francisco’s emergency services will be blocked by Waymos.

I’m very much in *favor* of line length limits in code. Personally, I still use 80. This isn’t some “stuck” thing—holding on to obsolete tech–the 80 character width was in fact chosen, way back, for readability reasons. (I’ll have to find citations but I remember this) I might be convinced to use 100 or 120, but there must at least be a limit.

The rationale is code readability, especially in the context of safety. If you cannot see all the code for the context you’re auditing, it is not safe.

Things I remember from the 1980s, the latest right-wing things-were-better-then craze, in London:

* You would have black snot at the end of the day if you spent much time in busy streets, especially central, especially if you took the underground.
* Every landmark was a dark slime color, from leaded gas/petrol painting itself over *everything*. I'm not kidding! The fabulous colors of the Houses of Parliament you see today were amazing, when it got cleaned.
* Everyone was poor and wanted to move

@ryanc Wonder how small that could get without all the calling convention overhead, which looks like about half the binary.

@regehr There's an insidious problem on targets where the ISA has a few instructions which rely on alignment (e.g some SSE/AVX and even recent ARM). Unaligned access works until the optimizer figures out a way to save a cycle by using the faster vector op, and then it breaks. This turns into a heisenbug where any attempt to track it down breaks the optimization.

@mekkaokereke This would mean that the largest fund managers voted against the proposal. This is interesting, because that means even the people in charge of a large fraction of the world’s wealth think DEI is a good idea.

Individuals and employees have zero (rounding error) influence in shareholder votes. The sum of all Costco insiders is something like 0.2%, and that’s likely almost all C-suite.

@ryanc I imagine it's like this (sadly) common setup:

* Receiver is a for() loop, either pulling bytes out of a UART, or executing a command.
* UART is either a 1 byte FIFO, or emulated GPIO.
* While executing a command, UART is dropped.
* Accessory, vice versa situation.
* This all "works" because the protocol will have a "documented" AC timing spec: T_inter_command_gap>=20ms.

There's nonsense like this common in devices everywhere, but to be fair it sure makes things cheaper.

@ryanc Rule of thumb is the firmware in any large appliance was written by a hardware engineer, who generally stop at “it works now”.

This reminded me of working on ID over HDMI, which uses I2C. Turns out for a large fraction of devices, if you spam I2C, e.g waiting for boot, their single work loop never gets around to doing anything other than I2C, including booting. You need to leave arbitrary gaps like you did there :(

@samir @ryanc
return 0 is it tho -0;

@fivetonsflax It also strongly attempts to imply the entire conflict is the fault of one person, which is such baby thinking.

The GOP, and largely the whole of the nebulous fascist movement, is approaching a deadline when they likely cease to be viable through democratic means, so of course everything is getting dialed up.

Especially when the consequences are after the king is installed.

@mjg59 Some ISPs have magic TTL values which they use as an obfuscated way to route traffic differently (e.g no caps). Could be it’s accidentally tripping into that.

@ryanc Unwanted entity in the middle of an exchange, claiming credit for what each person is saying but introducing something bad in the process, and potentially taking it over. Fairly accurate.