eb0la

I have the same feeling, but after reading this report I believe big hyperscalers will survive the bubble when it pops.

Now I've got the feeling they don't have huge amounts of GPUs sitting in their DCs, but rented for Opex. In case the bubble pops they might get it at discount as CapEx (like Amazon did with dark fiber after the dotcom bubble).

Several european countries will be governed by fascist parties, fueled by social media. There will be no EUexit because they will discover how to use the EU to approve legislation that won't be popular.

Tourism in Mediterranean zone will plummet: it will be just too hot for tourist to enjoy during summer.

Food prices will soar everywhere because climate change will make harder and more expensive to produce vegetables.

Be careful about the culture. Higher management had a massive panic because "AI stopped working" in a product we make. They started an all-hands meeting. Everything worked except the AI responses. Turns out we just ran out of credits. The bad part is we told everybody we would run out of credits in less that 3 days. Nobody cared until we had production outage. Now procurement is complaining about token usage.

Before reading the article I was surprised to find them similar to old german Sout backpacks. They are really sturdy and durable: your kid needs just one of them for all primary school (grundschule). They are explensive (not so much considering 3-6 years of continuous abuse by kids), but when the kid gets tired of it, some people put them on sale. I have one that I know was resold at least 2 times and it still in perfect shape... Great for airport travel, btw.

Right now it is impossible to order cost-optimized servers in Falkstein, Nuremberg, or Helsinki. No matter if it is x86 or ARM.

Looks like they had to buy _a lot_ of new servers and it was very expensive.

It was a great opportunity to name a unix tool "mcfly" or just "Marty" for time manipulation. Better luck next time, I guess.

Reminder: always travel with a spare phone with a broken screen just in case.

Keeping your sanity in tech is underrated.

This reminds me of Nortel Command Console back in 2000-2005!

I worked for a telco company that had a lot of Nortel Passport devices (does anyone know what Frame Relay is?). We started changing the network from Nortel to Cisco. Cisco used telnet (later SSH), but Nortel people were extremelly reluctant to switch.

Turns out the Nortel network managment system (nortel nms) had a very interesting feature: you could open the command console to connect to one of the passport devices... or you could connect to a device group (or all the network) and run the same command in all devices.

This was great for auditing which version had every single device in the network... or for changing access-lists globally.

I wonder what's the output of running 'hinv' on Irix.

I need to recover my old Irix CDs and find out.

Probably the police wants to gather some evidence. The easiest way to block a wifi tap is to change the wifi password... or wifi name. Both things will be noticed by the eavesdropper because they will block the device *but* it is an easy way to block it asap.

What I would do:

1 - Get an old android phone. 2 - Configure the phone to act as an access point with the same wifi name and password. 3 - Change the wifi password or name in the router. 4 - Change wifi passwords in laptop, etc... (I know, it is a pain). 4 - Activate the phone access point. 5 - * IF * the listening device connects to the phone AP wifi, you know there's a wifi tap. You can connect the phone to your home wifi later... and it will look transparent to the stalker.

Best of luck with this situation.

MySQL or Postgres are the DB of choice if you want a managed database in the cloud.

Probably Postgres is there because you can use it as a queue (https://livebook.manning.com/book/just-use-postgres/chapter-...)

I cannot dísclose muy current compensation due to an NDA: salaries are company propietary information.

I am unable to dísclose that information.

I would pay if I can use Clip Studio Paint without lag. In fact, I will try another time this easter. If works, I will need to donate.

One of the reasons behind the dollar strength is that the US has a huge population.

Even if the central bank might does a bad job and make a mess of the economy, the activity of 350 million people is hard to ignore.

Is it enough to _fully_ sustain the US dollar?

Who knows, but at least there is a floor, even if everybody stopped using US dollars for international trade.

Not just Amazon - I guess the Oil and gas industry is now run on the cloud. They used to have big SGI machines 30 years ago... but I bet everything is on the cloud now using GPUs.

If you have to walk the dog and know in advance it will stop or start raining in 20 minutes....

For me the only reason to keep windows 11 is Clip Studio Paint doesn't work yet in Linux... I use it mostly on my Android tablet BUT I want to have an alternative in case something is easier on a computer

But in the other hand you don't have to worry about mass shootings. You can freely walk (mostly) wherever you want without risking your life (that is not normal in most of the world). And you're not going bankrupt because of a minor/medium medical condition.

Europe is a _very_ different place.

Not everything here is so bad.

Politicians use it a lot. Because media and journalist started using it.

Until now nobody thought it was a problem. At least not a big one. The EU made some moves to define a "cloud computing" platform for Europe, and very little people paid attention because business-wise it was very difficult to compete with US corporations that have vast amounts of money in cash and find easy to get funding.

But now there are some (small) alternatives.

LIDL has its own cloud for retail.

And I believe T-Systems sells some cloud computing for goverments based on OpenStack...

Small steps, but steps.

In Spain you need to be ENS-certified (esquema nacional de seguridad) in order to provide services to the goverment. Nowadays it is similar / aligned to NIS2 certification.

But you need to certify more than just apps. Processes are more important than apps.

We started with PGVector just because we already knew Postgres and it was easy to hand over to the operations people.

After some time we noticed a semi-structured field in the prompt had a 100% match with the content needed to process the prompt.

Turns out operators started puting tags both in the input and the documents that needed to match on every use case (not much, about 50 docs).

Now we look for the field first and put the corresponding file in the prompt, then we look for matches in the database using the embedding.

85% of the time we don't need the vectordb.

It is true that (most) people don't care, but... ... if you're beign charged by RAM (CPU is essentially free) ... ... AND you develop an application that's also your source of income ... THEN memory usage matters.

At least the first years.

I believe we're in the middle of a "perfect storm" and AI is not the only one to blame:

- Right now A LOT of PCs are getting out of date because Windows 11 wants some new hardware that's missing in older PCs. - At the same time, smartphones were starting to use and need more memory. - Modern cars (mostly electric) have much more electronics inside than older ones... they're now distributed systems with several CPUs working together along a bus. - The cloud needs to upgrade and newer servers have much more memory than older ones with almost the same foorprint (wich means you need to lease less datacenter space). - And GPUs for AI are in demand and need RAM.

But only AI is to blame although we're living in a perfect storm for RAM demand.

Yes: Orange clown very bad.

That's the point: how can you tell WHEN you are going to reach a place you've never been before traveling an uncertain path?

Making mistakes over and over again. And adding a lot of time buffers just to be safe

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.