Veserv

44 ms·

481.
by Veserv 2y ago
The entire article is arguing that Tesla has presented insufficient auditable evidence to support their conclusions and separately has done so to deceive and defraud their customers. Are you disagreeing with the first or second conclusion?
482.
by Veserv 2y ago
That is a fundamental misunderstanding of the reliability level needed for fully autonomous vehicle operation. A fully self driving vehicle operating outside of a testing safety protocol in use by consumers requires a average disengagement
483.
by Veserv 2y ago
This is not at all comparable to inline assembly which interleaves two different languages into the same source file. The presented examples are just a straight distinct assembly language associated with the golang ecosystem used in their o
484.
by Veserv 2y ago
To add-on, utility scale solar generation is the majority of generation and has consistently outpaced small scale generation in growth. Starting at 11,233 small scale out of 26,482 in 2014, ~42.5%, versus 73,406 out of 236,090 in 2023, ~31%
485.
by Veserv 2y ago
You hook the kernel write. That is why I said hook the relevant kernel operations. The primary complexity is actually in creating a consistent timeline with respect to parallel asynchronous writes. Record-Replay systems like rr usually just
486.
by Veserv 2y ago
That does not seem to be correct. The documentation indicates APC [1] can only occur from a waiting/blocking state. So, the program is in a consistent state and can only be on a few known instructions, unlike signals. As such, most fun
487.
by Veserv 2y ago
You are correct. A time travel debugging solution that supports recording the relevant system call side effects would handle this. In fact, this system call is likely just rewriting the program counter register and maybe a few others, so it
488.
by Veserv 2y ago
SI prefix words are just kind of silly. We should just use the exponent as a number instead of having a different word for every 3 zeros. 10.7 E15 Wh or something similar. Scales to everything, you do not need to know any mapping, and direc
489.
by Veserv 2y ago
Because if you integrate with Apple Pay or Google Wallet then iOS or Android become the platform that delivers tailored applications, not your "everything app". Almost like that is exactly how it currently works in much of the wor
490.
by Veserv 2y ago
Yes, very few countries have a app where you can do "everything" in it. "This clear definition is really hard to achieve, therefore it must mean something different" is not a counter-argument. I observe that these applic
491.
by Veserv 2y ago
India and Brazil [1]. Only their two biggest markets [2]. [1] https://faq.whatsapp.com/1293279751500598 [2] https://worldpopulationreview.com/country-rankings/whatsapp-...
492.
by Veserv 2y ago
WhatsApp is not a "everything" platform in most of Europe. Do you open WhatsApp to play music, games, order takeout, order online, pay for rent? It is just a very popular chat application in most of Europe in contrast to a "e
493.
by Veserv 2y ago
Payments. The sticky part of "everything" app is the payments platform. With a payments platform you have a internal "funded" ecosystem. Anybody who can not take payments from the "everything" app has a incenti
494.
by Veserv 2y ago
You do realize the Tesla "patent pledge" [1] is one of the most unfair, unethical, and downright despicable legal contracts to ever attempt to cynically claim the moral high ground? To conform to it you must not have: "assert
495.
by Veserv 2y ago
There are really two things at play here. First, every manufacturer should be required to collect and produce scientifically rigorous evidence of safety for these systems. Any system that does not increase danger, i.e. within the range of h
496.
by Veserv 2y ago
Signal handling is distinct from a different thread because it blocks the execution of the “preempted thread” until the signal handler completes. As a example, if the preempted code grabs a lock for a resource, then signal handler completio
497.
by Veserv 2y ago
The relationship between the occurrence of a past event and the second it occurred can not be changed, barring relativity. As such, storing the absolute second offset unambiguously persists the actual time the event occurred regardless of a
498.
by Veserv 2y ago
It is important to note that this is only relevant for future events where legal/civil time, a means for squishy humans to coordinate actions in the future, is what matters. For instances like “bake my bread for one hour” you care abou
499.
by Veserv 2y ago
Really, you are arguing that I get to kill 1645 in the UK before I have to show any scientifically rigorous evidence? I get to do 1645 drives and kill the passenger every time before anybody cares? That is not only insane, but empirically f
500.
by Veserv 2y ago
On a normalized basis in the USA? 1 fatality per ~4000-5000 years of driving. 1 injury per ~60-70 years of driving. Can you point to scientifically rigorous evidence, at least reaching the bare minimum of fit for publication, of the absolut
501.
by Veserv 2y ago
Well of course. In the USA, health insurance is legally required to spend at least 80% of premiums on medical care [1]. So, if they paid for only 50$ of care, they would only be able to make at most 62.50$ of revenue. The only way to keep r
502.
by Veserv 2y ago
Yep, every actually efficient logging system does it that way. It is the only way you can log fast enough to saturate memory bandwidth or output billions of logs per core-second. You can see a fairly general explanation of the concept here:
503.
by Veserv 2y ago
Okay, I should have said “a” instead of using “the” when there is no clear antecedent allowing it to be interpreted ambiguously in exactly that single sentence which apparently invalidated the fact that I was obviously talking in generaliti
504.
by Veserv 2y ago
Sure, but the article is talking about grep, not write() or libc implementations. grep buffers writes with no flush timeout resulting in the problem in the article. grep should probably not suffer from the problem and can use a write primit
505.
by Veserv 2y ago
You are not meaningfully sacrificing throughput because the timeout only occurs when you are not writing enough data; you have no throughput to sacrifice. The threshold and timeout should be chosen such that high throughput cases hit the th
506.
by Veserv 2y ago
You will generally only stall indefinitely if you are waiting for new data. So, you will actually handle almost every use case if your blocking read/wait also respects the timeout and does the flush on your behalf. Basically, do it syn
507.
by Veserv 2y ago
You do not schedule a timeout on each buffered write. You only schedule one timeout on the transition from empty to non-empty that is retired either when the timeout occurs or when you threshold flush (you may choose to not clear on thresho
508.
by Veserv 2y ago
The solution is that buffered accesses should almost always flush after a threshold number of bytes or after a period of time if there is at least one byte, “threshold or timeout”. This is pretty common in hardware interfaces to solve simil
509.
by Veserv 2y ago
This always comes up. A “recall” is not a description of the remediation, it is a description of the problem. A recall is a public dangerous defect notice. The dangerous product version can no longer be deployed, existing systems suffering
510.
by Veserv 2y ago
Ah yes: "Don't you understand, the problems in this field are harder. That is why the methodologies that easier fields reject as being grossly inadequate for their easier problems will work on much harder problems." Even igno

More ›