xmddmx

^F Latency - no hits found.

Latency is one metric by which computing has gotten noticeably worse over the years.

Daily, I encounter multiple situations in which I take some action (click a button, type text in a text field, etc.) and the response is visibly delayed - from several hundred to several thousand milliseconds.

This used to happen only on websites, but lately is also happening on apps installed on my computer. The cause? My theory is that everything is now some sort of monstrosity of React or similar web app, running inside a browser process, and possibly even that local apps are sending data round-trip up to the mothership for tracking, surveillance, synchronization, etc.

And the UX is terrible - often while this is happening, stale data is showing, or there is no visual indication a process is underway.

Agree. And this is all made worse by the fact that there seems to be a bug in iCloud keychain syncing - I can change the password on my mac, verify it works, but when I go over to iPhone or iPad, the old password is still being used. Not sure if there's a way to force the sync process on both ends?

Sure, but it also took them 7 hours to email their own customers on the email infrastructure run by DreamHost themselves. Since they run the IMAP and SMTP servers, seems like they could have done that in seven seconds.

Of course, if they are nuking everyone's email password that doesn't help for notification purposes -- but as of this morning, all my email logins are still working fine, so it seems they didn't actually do that? Confusing.

Super displeased with DreamHost's handling of this. I have multiple master accounts. One of them got the email around 12pm today (over 48 hours after the incident started) saying...

"As a precaution, we have expired the passwords for some email and FTP/SFTP users on your DreamHost account. Their old passwords no longer work, and no new password has been set."

But none of those email passwords were actually expired - all were working. This is confusing - did they only expire passwords that were under threat? If a password wasn't expired, should it be trusted?

The same email suggests you change all email passwords you control.

Another account received no email warning from DreamHost so I assumed it was fine, but 7 hours later it also received the same generic email.

My complaints:

- sucks that it took them 48 hours to identify the breach - perhaps understandable

- but once the breach was identified, how the fuck does it take them over 7 rolling hours to email their customers? That's terrible.

- confusing messaging: DreamHost are claiming that passwords are expired, but in testing multiple accounts, all of them are still working. Do I need to rotate these? (I probably will just for safety).

I'm a 25+ year DreamHost customer and generally satisfied, but this feels like amateur hour.

What quant and engine are you using? The smallest ones I’m seeing on ollama are > 100GB e.g. https://ollama.com/library/qwen3.8-flash-next/tags

If only Apple computer had employees that would fix this system (that they are 100% responsible for building) :-)

Even if Apple is too cheap to pay the $0.00001 to get me real time temperature updates, you would think there would be an OS API to query the temperature, and that the various apps and complications would query this to get the source of truth.

But from the actual UI/UX, it must be that every App, and complication, and UI element are caching their own value.

That's the only way I can explain seeing 3 different temperatures displayed within a few seconds.

(NB: these are all Apple apps, I'm not using any third party apps)

Right, that seems to be what is happening.

The question I'm asking is "why" - why would Apple ship code that is so horribly broken (from a UI/UX) perspective.

I thought for sure that this slow-temperature-update bug would be fixed in WatchOS 27, but it's just as bad as in 26 and before.

Sometimes the temperature UI will show "--" instead of a temperature, until it finally updates, so clearly they do have code for the "no temperature data" case.

It just boggles the mind that "6 hour stale temperature data" doesn't trigger this same behavior.

I've seen bugs like this which have led me to wonder if apple has a problem with generating UUIDs which aren't really unique.

It's the only way I can explain some of the totally bizarre Siri (and other iOS and macOS) bugs.

Agree - I've had almost all of these problems with Siri.

Another thing I've noticed is the horrible lag with updating the UI. An example: On Apple Watch I have the temperature complication bottom center. It sometimes shows the current temperature, sometimes it hasn't updated in minutes or hours (bug #1). For example right now, it shows 66 which I know is incorrect. I tap it to go to the Weather app, a different temperature will be shown (56) (bug #2) and then about 10 seconds later (bug #3) it finally updates to the actual temperature (72).

If I go back to the home screen, finally the complication shows the correct temperature (72).

These kind of async refresh bugs are not unique to WatchOS, it's endemic throughout the Apple ecosystem.

Why? Is this poorly coded JavaScript web view apps? Is this React? Some sort of other async framework that works fine when you are testing on the Apple campus but doesn't work in the real world?

Would running the unit in Fahrenheit improve the sensitivity?

Even more confusing. Do we know which data is in which keychain?

Disturbing if true, as it suggests my "everything" backups aren't really full backups at all.

I wonder, if you have enabled iCloud keychain, does that provide another way to get your passwords back?

Suppose my MacBook is stolen, but I still have my iPhone. Could I use the passwords app on the iPhone to retrieve my passwords. Would that include all of them or only a subset?

Another way (which happens to be correct) to say tenth of a second is 100 000 000 (one hundred million) nanoseconds. You were off by a factor of 1000!

There's a meta-level of irony here that's important to note.

TFA is defending the use of AI, and it very clearly (to me) used AI to analyze the data and present the results.

In doing so, the author used statistics in a way they do not appear to understand, and ended up making numerous false claims (you can see the thread discussing these here https://news.ycombinator.com/item?id=48417626 )

In short, the study doesn't have sufficient statistical power, and is making "no difference" claims that aren't justified.

The meta-irony is this: the author used an LLM to interpret data in this study, and seems to have made the same category of mistake (confidently asserting falsehoods) that the study was supposed to be investigating (confidently submitting bad commits to the rsync project).

You say "the underpowered study did not find evidence". Not true, it found quite a bit of evidence - many statistics were presented. There is no absence of evidence. The author wrote about the evidence, presenting P values and other statistics.

Of course the critical part is not the numbers, but what they mean.

So, what does the evidence mean?

The author interprets it to mean that there is no difference. They state this several times:

"46% EXACT PERMUTATION TEST P-VALUE (ONE-SIDED, H₁: CLAUDE MEAN > HISTORICAL)[...] What this p-value tells us is There's nothing unusual about the Claude group."

"74% ONE-SIDED P-VALUE (H₁: CLAUDE MORE LIKELY ABOVE MEDIAN) Fisher's exact test asks: if we split all releases at the historical median (0.74 sev/10c), are these Claude releases significantly buggy than previous releases (more likely to land above the median)? With a p-value of 74%, the answer is a decisive no. "

In an under-powered study, when a P value is above your alpha level cutoff (.05, .01, whatever was chosen) you can't distinguish between "no effect" and "could be an effect, but I didn't see one".

Perhaps in an “everyday language” way, but not in the technical, statistical sense.

In an underpowered statistical study, a claim that two experimental conditions did not differ are not persuasive.

The concept you need here is "Statistical Power".

The ELI5 version is that there are two mistakes you can make when looking at a P value:

Type I error, where your P value is falsely low. In the experiment being discussed here, it would lead one to conclude that AI code is worse. Otherwise known as a false positive.

Type II error, where your P value is falsely high, leading you to conclude that AI code is no different. Otherwise known as a false negative.

https://en.wikipedia.org/wiki/Power_(statistics)

One can calculate statistical power for a given experimental protocol.

My hunch is that if you did this, you would find this experiment is grossly under-powered.

This means you can't make the "absence of evidence" claim.

On a M4 Pro MacBook Pro with 48GB RAM I did this test:

ollama run $model "calculate fibonacci numbers in a one-line bash script" --verbose

  Model                         PromptEvalRate EvalRate
  ------------------------------------------------------
  qwen3.5:35b-a3b-q4_K_M         6.6            30.0
  qwen3.5:35b-a3b-nvfp4         13.2            66.5
  qwen3.5:35b-a3b-int4          59.4            84.4

I can't comment on the quality differences (if any) between these three.

See my other note [1] about bugs in Ollama with Qwen3.5.

I just tried this (Ollama macOS 0.17.4, qwen3.5:35b-a3b-q4_K_M) on a M4 Pro, and it did fine:

[Thought for 50.0 seconds]

1. potato 2. potato [...] 100. potato

In other words, it did great.

I think 50 seconds of thinking beforehand was perhaps excessive?

[1] https://news.ycombinator.com/item?id=47202082

See my other note about bugs in Ollama with Qwen3.5.

I just tried this (Ollama macOS 0.17.4, qwen3.5:35b-a3b-q4_K_M) on a M4 Pro, and it did fine:

[Thought for 50.0 seconds]

1. potato 2. potato [...] 100. potato

In other words, it did great.

I think 50 seconds of thinking beforehand was perhaps excessive?

Ollama users: there are notable bugs with ollama and Qwen3.5 so don't let your first impression be the last.

Theory is that some of the model parameters aren't set properly and this encourages endless looping behavior when run under ollama:

https://github.com/ollama/ollama/issues?q=is%3Aissue%20state... (a bunch of them)

> Sheldon Brown, a beloved iconoclast bicycle tech guru, died Sunday from a heart attack. He was 68 63.

Curious, what does "He was 68 63" mean. Is it a bicycle gear joke about his age at death?

On Mac Safari, holding shift and using the magic mouse to scroll up or down reverses the zoom direction.

This is both right (Shift-X is the reverse of X due to convention) But is also wrong (Shift-Scroll is the macOS gesture for scrolling on maps where Scroll alone doesn't zoom in or out).

TLDR: I really wish Apple would adopt the "scroll up to zoom in" convention used by the rest of the free world.

My memory is that it was named "Aero Glass" which heightens the irony of "Liquid Glass" sucking.

But I see many references to it being called just "Aero", but some call it "Aero Glass" [1]

Does anyone know the truth?

[1] https://www.pcmag.com/archive/rip-aero-glass-windows-8-stick...

Right, this annoyed me too - it was stated w/o attribution as if novel.

What is the name of the law when someone writes a think piece of "stuff I've learned" and fails to cite any of it to existing knowledge?

Makes me wonder if (A) they do know it's not their idea, but they are just cool with plagiarism or (B) they don't know it's not their idea.

You mean 1990. Someone graduating college in 1990 would have been about 21. That was 35 years ago, so they would be about 56 in 2025.

Math is hard.

The CFR is pretty clear, and I have experience with this (being both an IRB reviewer, faculty member, and researcher). When it says "is exempt" it means "is exempt".

Imagine otherwise: a teacher who wants change their final exam from a 50 item Scantron using A-D choices, to a 50 item Scantron using A-E choices, because they think having 5 choices per item is better than 4, would need to ask for IRB approval. That's not feasible, and is not what happens in the real world of academia.

It is true that local IRBs may try to add additional rules, but the NU policy you quote talks about "studies". Most IRBs would disagree that "professor playing around with grading procedures and policies" constitutes a "study".

It would be presumed exempted.

Are you a teacher or a student? If you are a teacher, you have wide latitude that a student researcher does not.

Also, if you are a teacher, doing "research about your teaching style", that's exempted.

By contrast, if you are a student, or a teacher "doing research" that's probably not exempt and must go through IRB.

Turns out that under the USA Code of Federal Regulations, there's a pretty big exemption to IRB for research on pedagogy:

CFR 46.104 (Exempt Research):

46.104.d.1 "Research, conducted in established or commonly accepted educational settings, that specifically involves normal educational practices that are not likely to adversely impact students' opportunity to learn required educational content or the assessment of educators who provide instruction. This includes most research on regular and special education instructional strategies, and research on the effectiveness of or the comparison among instructional techniques, curricula, or classroom management methods."

https://www.ecfr.gov/current/title-45/subtitle-A/subchapter-...

So while this may have been a dick move by the instructors, it was probably legal.

I was impressed by the lack of dominance of Thunderbolt:

"Next I tested llama.cpp running AI models over 2.5 gigabit Ethernet versus Thunderbolt 5"

Results from that graph showed only a ~10% benefit from TB5 vs. Ethernet.

Note: The M3 studios support 10Gbps ethernet, but that wasn't tested. Instead it was tested using 2.5Gbps ethernet.

If 2.5G ethernet was only 10% slower than TB, how would 10G Ethernet have fared?

Also, TB5 has to be wired so that every CPU is connected to every other over TB, limiting you to 4 macs.

By comparison, with Ethernet, you could use a hub & spoke configuration with a Ethernet switch, theoretically letting you use more than 4 CPUs.