bcrl

Postfix has a far better security track record. Exim has a steady stream of CVEs.

Personally, I find the predictable network names feature infuriating 100% of the time. Changing the name of network interfaces made nothing better, and actively made things worse for systems with a single NIC (virtually none of the systems I use with a single nic use the same name). Network interface configurations could already be bound to MAC addresses rather than names, and that had been implemented even before the enpXXX style names came into being.

Persistent ethX names were far better to me, and we had those before enpXXX via udev. At least then when I logged into a random system with a single NIC it was called eth0. With 2 NICs they were eth0 and eth1. Simple to predict, no thinking required (unless the system already had bound those names to other MAC addresses).

Usability dies by a thousand cuts. Forcing needlessly complex behaviour onto people with simple use-cases is not an improvement.

All you need are NAND gates.

McDonalds actually seems to have learned to take latency seriously. When their touch screen ordering systems were first deployed, the delay between tapping on an item or button was quite noticeable. These days the systems respond nearly instantaneously. I'm very glad there are people inside such a large organization that pay attention to that aspect of usability.

Now if only every other website on the internet would learn that latency matters...

BitKeeper tried to do that. Git was built because the commercial license of BitKeeper became unworkable for the Linux kernel community.

"Those who cannot remember the past are condemned to repeat it".

There were more ~26+ years ago. gcc and egcs had some subtle register allocator bugs that would get tripped up under heavy register pressure on i386 that were the bane of my existence as a kernel developer at the time.

Please name a computer science program that has an ethics component.

Yes, I wish software developers were more like actual engineers in this regard.

Software was already far down the bloat path by the time the Core 2 Duo came out, so the upgrade didn't make all that much of a difference in feel given how much latency was caused by software performing random reads off a disk. That's why SSDs made such a huge difference.

Back in the MS-DOS days, the amount of data needed to be read off a disk while the OS booted was negligible, so a second or two on a fast 486 felt amazing compared to the incredibly slow grind of watching code execute on an 8086 or slow 80286. Software was still in the space of having to run tolerably on an 8086, so the added resources of a newer faster machine actually did improve the feel of the system.

Many of the merge lanes in California are insanely short compared to those in the rest of the world. The worst are the ones that have merge immediately before an overpass and exit immediately after where merging and exiting have about the width of the overpass to change lanes. I found those infuriating when I used to visit friends in the Bay area. The pattern where I live is the opposite (long exit lane before the overpass and a long merge lane after) and provides far better margins of safety.

It's plausible that the AI companies have given up storing data for training runs and just stream it off the Internet directly now. It's probably cheaper to stream than buying more SSDs and HDDs from a supply constrained supply chain at this point.

Given the tectonic shift in priorities for Linux kernel development over the past decade, I'm willing to bet that many key developers would be open to a microkernel architecture now than ~25+ years ago. CPUs now have hardware features that reduce the overhead of MMU context changes which gets rid of a significant part of the cost of having isolated address spaces to contain code. The Meltdown and Spectre attacks really forced the security issue to the point where major performance costs to improve security became acceptable in a way that was not the case in the '90s or '00s.

To make things even more confusing, the high-density floppy introduced on the Amiga 3000 stored 1760 KiB

Please read the article in full. The GPU die where all the computations occur and the majority of power is spent will remain on TSMC.

TSMC plans their A14 process to be in high volume production in 2028. It will include backside power delivery introduced in their A14 process (expected 2026/2027 high volume production), which means it will be quite competitive with Intel.

https://semiwiki.com/wikis/industry-wikis/tsmc-a14-process-t... https://semiwiki.com/wikis/industry-wikis/%F0%9F%A7%A0-tsmc-...

There's an older article at https://www.igorslab.de/en/350-watts-for-nvidias-new-top-of-... which shows the breakdown of power consumption for GPUs. The GPU die itself is only 230W of the entire power budget.

The entire sentence is even less enthusiastic:

"The GPU die will remain with TSMC, but portions of the I/O die are expected to leverage Intel's 18A or the planned 14A process slated for 2028, contingent on yield improvements."

Reading between the lines: Nvidia will most likely design a TSMC version of those I/O die portions in case Intel fails.

Intel has a decades long reputation of failing its attempted foundry customers. Whether or not Nvidia's ownership stake is sufficient to overcome the inertia within Intel that has resulted in those failures remains to be seen.

Thanks for publishing your blog! The articles are quite enlightening, and it's interesting to see how semiconductors evolved in the '70s, '80s and '90s. Having grown up in this time, I feel it was a great time to learn as one could understand an entire computer, but details like this were completely inaccessible back then. Keep up the good work knowing that it is appreciated!

A more personal question: is your reverse engineering work just a hobby or is it tied in with your day to day work?

Ah yes, the "things are bad; we shouldn't try to fix them" argument. That isn't a philosophy which I subscribe to. People should very much consider the ethical implications of releasing software they created to the general public.

2.4 GHz is unreliable for me these days due to interference from bluetooth headphones and hearing aids that other people are using. The issues tend to only show up during extended periods of video streaming, and having looked at a bunch of traffic captures over the holidays, it seems to be limited to certain streaming services sending very large bursts of traffic at extremely high rates (likely from servers with 100+ Gbps interfaces using TSO to reduce CPU usage). That makes me think that the regularly paced bluetooth interference from real time audio streams limits the maximum viable burst size of a 2.4 GHz wifi radio.

Yes, this happened a bunch more over the Christmas holiday when we had an extra 3 or 4 younger family members all listening to music and videos over their bluetooth ear buds and headphones, which made it much easier to track down as it was quite a rare intermittent failure with only a single bluetooth device being active.

Security for maps is basically impossible. Maps tend to have to be widely shared within government and engineering, and if you know what you're looking for, it's remarkably straightforward to find ways to access layers you would normally have to pay for. It's a consequence of the need to share data widely for a variety of purposes -- everything from zoning debates within a local county to maps for broadband funding across an entire country create a public need to share mapping information. Keys don't get revoked once projects end as that would result in all the previously published links becoming stale, which makes life harder for everyone doing research and planning new projects.

Moreover, university students in programs like architecture are given access to many map layers as part of the school's agreements with the organizations publishing the data. Without that access, students wouldn't be able to pick up the skills needed to do the work they will eventually be hired for. And if students can get data, then it's pretty much public.

Privacy is becoming (or already is) nearly impossible in the 21st century.

I basically asked my math and physics teachers in high school what the Fourier transform was, but none of them knew how to answer my questions (which were about digital signal processing -- modems were important things to us back in the early '90s). If I had to do it over again, I would have audited the local university's electrical engineering and math courses in evenings. The first time MIT ran 6002x online back in 2012, the course finally answered a lot of those questions when touching upon filters and bandwidth.

What measures are being taken to ensure that this model isn't used to lower the cost of fraudsters committing grandparent scams by mimicking the voices of grandchildren?

The old manual tools were extremely slow. Modern fibre splicers mean that a dozen fibres can be spliced in maybe a bit more half an hour, although cable prep cam take a significant amount of time depending on the cable type, number of cables and splice closure. Even more if you're using a ribbon splicer that fuses 12 fibres per burn.

Cable still has one thing going for it: it tends to be cheaper for sports. Watching hockey games online requires subscriptions to 3 different streaming services just to follow a single local team, which is ridiculous.

I would be fine if data centers paid the full cost of their existence, but that isn't what happens in our world.

Instead the cost of pollution is externalised and placed on the backs of humanity's children. That includes the pollution created by those datacentres running off fossil fuel generators because it was cheaper to use gas in the short term than to invest in solar capacity and storage that pays back over the long term. The pollution from building semiconductors in servers and GPUs that will likely have less than a 10 year lifespan in an AI data center as newer generations have lower operating cost. The cost of water being used for evaporative cooling being pulled from aquifers at a rate that is unsustainable because it's cheaper than deploying more expensive heat pumps in a desert climate.... and the pollution of the information on the internet from AI slop.

The short term gains from AI have a real world cost that most of us in the tech industry are isolated from. It is far from clear how to make this sustainable. The sums of money being thrown at AI will change the world forever.

I just had an incident like that a month and a half ago. Customer reported repeated internet outages in the morning. Lots of back and forth by phone and email, and the only conclusion I could come up with is that the ONU or router was flaky and needed to be replaced.

Nope. Ended up going on site one day. It turns out that the power bar everything was plugged into was sitting on the floor at the back of the desk in the customer's office. When they sat down first thing in the morning, they would often jostle one of the power supplies just enough to cause a restart. Moved the power bar over 2 feet to the left where feet couldn't reach it, and the problem was solved.

LibreOffice is good enough for many use cases. A competing product doesn't have to be a 100% match feature for feature to be Good Enough for most users.

I wonder if that sentence will have any discernible meaning 100 years from now.

Or it makes the many deal with the cost of cleaning up the garbage created by 1. Privatize the profits, socialize the losses.

Gnome considers features a bug, so not at all surprising.

It makes me wonder what weight is given to content from 4chan during llm training...

I think it comes standard!

The third party transport links feeding my company's network have suffered 2 lengthy outages due to backhoes hitting buried conduit with fibre installed. Underground outages tend to be the longest outages since repairs can be quite involved. One of the outages was due to a large municipal construction project (LRT) where the buried conduit's location had been properly marked and yet was still hit anyways. They managed to pull on the fibre hard enough that repairs extended to a total of 4 manholes since the splices in the adjactent manholes were completely destroyed.

The other more recent underground hit occurred due to a process failure since the locates were issued before the carrier had installed their conduit and fibre, with the end result being that the other party doing construction had an all-clear from a locate completed before the fibre was in the ground. Ooops!

Dump truck hits are far more common around Ottawa. I'm aware of at least 8 in the past 5 years around where our network exists, which is only covering about 30 km. Thankfully when they have hit the poles we have fibre on, our fibre / strand was not damaged. The incumbent's fibre on the lowest strand on the pole was not so lucky. By my estimates there must be at least 1 near miss by a dump truck per day in the city. It's not surprising given the abuse dump trucks take.

If we had more fibre installed I'd love to have a proper ring to make these physical outages non-service impacting.