eb0la

My fault: I just compiled the instruction with an assembler instead of looking up the actual instruction from documentation.

It makes much more sense: resetting ax, and bc (xor ax,ax ; xor bx,bx) will be 4 octets, DWORD aligned, and a bit faster to fetch by the x86 than the 3-octet version I wrote before.

You're right. I forgot that!

I remember a lot of code zeroing registrers, dating at least back from the IBM PC XT days (before the 80286).

If you decode the instruction, it makes sense to use XOR:

- mov ax, 0 - needs 4 bytes (66 b8 00 00) - xor ax,ax - needs 3 bytes (66 31 c0)

This extra byte in a machine with less than 1 Megabyte of memory did id matter.

In 386 processors it was also - mov eax,0 - needs 5 bytes (b8 00 00 00 00) - xor eax,eax - needs 2 bytes (31 c0)

Here Intel made the decision to use only 2 bytes. I bet this helps both the instruction decoder and (of course) saves more memory than the old 8086 instruction.

Every 3-4 years RAM prices spike. There is always an excuse like a fire in a factory. I believe the truth is 1) we have little amount of suppliers, and 2) supply is very near the limit of what can be sold.

I must confess I felt a lot of lust looking at the self color calibration feature.

It is extremely useful if your work ends up in paper. For photography (edit: film and broadcast, too) would be great.

My use case are comics and illustration, so a self-color-correcting cintiq or tablet would be great for me.

Actually power efficiency was a side effect of having a straightforward design in the first ARM processor. The BBC needed a cheap (but powerful) processor for the Acort computer and a RISC chip was When ARM started testing their processor, they found out it draw very little power...

... the rest is history.

Comics Not just reading, but drawing them. For me nothing is better at inking than a Pentel Pocket brushpen. I do some lino engraving from time to time as well.

Totally makes sense. If you go to any mobile operafor you will see a lot of huawei hardware. From CEP routers and switches to multichassis routers in datacenters, and whole cell phone tower equipment.

They own the nework, they will better know than anyone how to tap it.

I believe #1 is true; but not #2. It's just that those rocks are more common than we thought. And we thought they were uncommon because we weren't able to spot them... yet.

Glad to see Amoeba here. Last time I heard about it I was still in college X-D

I guess academia was too far from what industry needed. They where thinking about the future, and we needed more stable boxes and sometimes a cluster for critical loads.

All the microkernel stuff Tanenbaum defended makes a lot of sense if you have a distributed OS: you just need to talk with your "local" service with the API you have at hand... and it is the same everywhere (for instance, your SOAP server - which is the equivalent of etcd in K8s).

I've seen some ADA positions in the automotive industry as well.

edit: This comment better hightlights the reason -> https://news.ycombinator.com/item?id=44233851

WHERE are you located? In Europe there are a lot of companies that need to be NIS2-compliant but have little (or no) idea about it. A service offering might work... unless you want to retire before your 30th birthday.

How about the opposite? You're not a paying customer, don't use trials... but let us put your logo in our website as a valued user... with some praise from the CTO, please.

I agree 120% with you... ... but I am wondering about how good you are using free tiers. IMHO the free tier in cloud/saas just offsets the initial costs of using the cloud/saas. So... unless you're really small free tiers won't work for you.

I believe all options you suggest are more than OK, but. Why don't you limit the trial with some capacity limits? Say, 1000 vms for installation. Of course, you'll need to have two artifacts: one for paying customers, and a second one to non-paying ones.

I had a project related to this in a big telco about 10 years ago. We used Citrix because a lot of apps were not native (yet).

Women in tech loved the concept: instead of carrying a laptop they could use the medium-size tablet we provided and have a wireless mouse and keyboard both at the office and at home. We also provided a monitor for both home and the office.

The project didn't lauch commercially because in our tests the microusb to micro-hdmi cables caused a lot of issues and management decided against it.

IMHO Cook is following good development practices.

You need to know what you support. If you are going to change, it must be planned somehow.

I find Torwalds reckless by changing his development environment before release. If he really needs that computer to release the kernel, it must be stable one. Even better: it should be a VM (hosted somewhere) or part of a CI-CD pipeline.

Just write what you know: they won't be able to know where were you...

... but they will be able to know if you're lying to them because they are trained to read people.

Not just Microsoft. I remember iPlanet directory server... which got renamed to Sun Directory Server, SUN Java System Directory Server, Sun Directory Server Enterprise Edition, Java One Directory Server, Oracle Java Directory Server...

I believe Oracle called everything back to iPlanet because customers never used the new names.

I believe when we talk about this there's a big misunderstanding between Copyright, Trademarks, and Fair use.

Indy, with its logo, whiplash, and hat, is a trademark from Disney. I don't know the specific stuff; but if you sell a t-shirt with Indiana Jones, or you put the logo there... you might be sued due to trademark violation.

If you make copies of anything developed, sold, or licensed by Disney (movies, comics, books, etc) you'll have a copyright violation.

The issue we have with AI and LLM is that: - The models compress information and can make a lot of copies of it very cheaply. - Artist wages are quite low. Higher that what you'd pay OpenAI, but not enough to make a living even unless you're hired by a big company (like Marvel or DC) and they give you regular work ($100-120 for a cover, $50-80/page interior work. One page needs about one day to draw.) - AI used a lot of images from the internet to train models. Most of them were pirated. - And, of course, it is replacing low-paying jobs for artist.

Also, do not forget it might make verbatim copies of copyrighted art if the model just memorized the picture / text.

If you're paying for tokens used to generate that...

My father was born in a small village in Guadalajara, Spain. I remember in the village my grandma and other neighbours tool their chair outside their homes to talk at the end of the day. It is great to see good things coming again. Do it... MORE.

If that happens, count with me to use Anubis to factor large primes or whatever science needs as a background task.

First time you go to London, you visit everything else; but Greenwich is great. Probably because is far from city center, but totally worth the visit. I guess next time I travel to London I'll stay there.

If visit Greenwich Observatory you'll see Harrison sea clocks H1, H2, and H3.

I had the luck (and honor!) to meet a volunteer guide - Stephen - that knew everything about clocks and explained it perfectly.

When you meet him, say hello :-). My kid barely spoke English and Stephen had the patience and virtue to answer everything with a smile.

Best guide I ever met.

Excalidraw and PlantUML Each one has its own benefits / drawbacks. PlantUML is great for version-controlled files. Excalidraw for throwaway diagrams and discussion. Ocassionally I draw comic strips with Excalidraw.

A lot of old hardware uses TFTP and RARP to boot. RARP just will get you the ip address, and the rest is hardcoded somehow in the machine - needs very little memory on boot. For BOOTP you need some intelligence to know where are your files. TFTP is also cheap in memory to use. UDP without flow, no nothing. Just send me the next packet in sequence when I ask you to do so.

I remember having trouble some years ago upgrading old Cisco routers because the image was bigger than what TFTP can handle.

Just a question: if you're adding stuff at the end of the query... does the previous steps get cached?

I mean, if you type first: FROM customers |> WHERE signup_date >= '2024-01-01T00:00:00Z' And then you add (and execute) |> COUNT() as tally GROUP BY CHANNEL

Does the first expression get cached by BQ?

First thing I thought is "this is a SELECT * in disguise", which is something that you SHOULD never do in BigQuery. If you can combine it with CTEs, seems good because it adds legibility for complex queries. Also looks easier for the planner. Not for the optimizer, though.

Lawyer-up. Get proper legal advice.

I won't pay anything that has to go to the IRS or similar to a corporation.

I might be OK for the company, but might put you in a big trouble with the goverment. Even if they act in good faith, you will need to do some paperwork or it will trigger an audit with the IRS that will take _your_ time, not theirs.