Meh... my BE 19000 just popped this up:

```
Authorize Third-Party Services
Once enabled, we will share your clients' information to a third-party services to identify your clients better. We won't save your private information.
```

Like I really believe that...

We made the mistake of buying a TP-LInk Onada switch for the office.

We use VLANs.

We setup the VLANs.

The switch immediately began sending VLAN 1 traffic to VLAN 2 and vice-versa. Apparently they expect you to setup L3 routing to prevent this? Doesn't this sound like a device which should be recalled and the company fined heavily for falsely claiming VLAN compliant?

Imagine you doing this and you rely on the L3 routing for the VLAN seperation. What about non-IP traffic? What about pretending to be on a different IP network? Does it move you to the other VLAN no matter what port you are on?

We sent it back and got a TRENDnet instead.

Took us most of the day to figure out the switch was mixing VLAN traffic between VLANs as if they were all one single VLAN.

If we could figure out how to charge them for false product and associated costs to recover from it?

Bigger impact than you think. Not everyone has updated to the buggy versions yet in their various dev cycles.

A new bug was filed just last week where when they updated the Java to run in their Windows Containers: https://github.com/adoptium/ad...

16 Hours ago the was the comment added: We have several customers affected by this issue on an eclipse RCP based application + RDP. on the ticket https://github.com/adoptium/ad....

Note, these issues are getting file via github and not openjdk site.... that appears to be invite only.

And no this is not a "minor" bug. It was causing absolute havoc at my office where we are constantly connecting and disconnecting to various systems during the day via RDP to startup and monitor various tasks that require human interaction. If `tech a` connects to `tech b`'s session, the session switch can cause the headless exception as well.

Issues like this don't help. A *breaking* change introduced in a minor dot release?

Java + Windows RDP disconnect causes Java Headless Exception

> Starting with Java 17.0.12 and 21.0.4 Java fails with a Headless exception when running a full GUI app via RDP and the RDP session is disconnected. The remote desktop is running and can be reconnected with. All non-Java apps are still running and updating their UI's just fine. It is only Java GUI apps that are crashing.

> If the Java program attempts to do GUI operations such as create a modal progress bar while the RDP session is disconnected, a Headless exception is thrown, even though the environment is not headless, just disconnected. ...

At my office we use XCP-ng, though it lacks support for nested virtualization with Windows guests, which breaks WSL2, Docker on Windows, etc. They are supposed to be working on a fix. See [https://github.com/xcp-ng/xcp/issues/105](https://github.com/xcp-ng/xcp/issues/105).

Generally, setup is easy, clustering is a thing, you can live migrate VMs to other hosts, and you can live migrate VM disks to other storage arrays as well. There is also support for vTPM in the upcoming 8.3 release and it the current LTS 8.2 supports secure boot.

AT&T should hire staff and fix the nested virtualization issue and release their work as part of the open source setup. Maybe REALLY screw over Broadcom that way.

-Mike/NewsRx

So... no sudden magic creation of 6502 and 6510 microprocessors from nothing

homophobic much?

Do to inaction with vaccines, are we at the cusp of another form of Smallpox type disease getting out everywhere?

Smallpox is *EVIL*. A thousand times worse than the hemorrhagic influenza of the early 1900's.

There is a reason so much money was spent on eradicating smallpox.

Now we have a replacement candidate that seems to be only of concern to the WHO?

Ever hear of airplanes? boats? asymptomatic vectors?

It is also spread by touch ANY TOUCH Think doorknobs, sit down toilets, et al.

It is not a sex disease, sex just happens to involve a lot more touch.

-the dog

> The maximum air temperatures this week — generally between 105 and 115 (41 and 46 C) — have only been somewhat above normal. But the dew points — which are a measure of humidity — have been excessive, climbing well into the 80s (27 to 32 C). In the United States, any dew point over 70 degrees (21 C) is considered uncomfortably humid. It's the very high dew points that have propelled heat indexes up to 30 degrees (16 C) above actual air temperatures.

AI generated? Or do we now measure humidity using the temperature scale? What other AI generated issues might be hiding in here?

dew point != wet bulb temperature

heat index != wet bulb temperature

humidity of 80s is percents not F

Interesting read (the paper).

a) I wonder how it would perform on a classical CISC CPU such as the x86_64 vs classical RISC CPU such as Arm? (Apple may have a real leg up here?)

b) How hard would it be to take an existing model, such as Llama-3-8b-Instruct, and "quantize" the weights to be compatible with the non-MatMul architecture?

c) Is the GPU market going to collapse in 6-months after someone releases a new Llama-nonMatMul or Mixtral-nonMatMul model using the new architecture? Or does a GPU still have the advantage here? (I haven't done low-level machine level programming in decades... not sure how SIMD, et. al., relate to all this).

Really sounds like a setup for HIPAA and other private not legal to be shared stuff ending up on servers in jurisdictions not authorized for receipt of such materials.

Think Europe and their data policy if someone installs windows 11 and it silently starts shipping stuff to a non-compliant backup site in the US...

Medical workstation nightmare. If it is enabled on setup, what is to prevent the later activation by automated means not initiated by the user with the user being unaware? Who becomes responsible for the potential breaches of such protected data?

Only $6G/shot, twice yearly...

We use XCP-ng in the office, but it has one major flaw. You can't nest M$ Hyper-V in it. The M$ Hyper-V fails under XCP-ng, but seems to work in other nest virtualization setups. Long running bug that really needs fixing, as it prevents being able to run WSL2 on virtualized workstations. And I understand it causes issues with Docker on windows as well.

https://github.com/xcp-ng/xcp/...

So... the model weights being published on huggingface don't count for anything? (https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1)

Also, the blog clearly states: Licensed under Apache 2.0.

Isn't there some way this can be turned into a class action lawsuit treating it as a fully anti-competitive action?