I used to agree until I found Level One Tech. I use this, and I'm very happy switching my two DP monitors, and a USB hub with mic, speaker, camera, etc. Its supports both USB 3.0 and DisplayPort 1.4a and use it to switch between my work system and my gaming PC.

https://www.store.level1techs....

and it rankles that I have to pay any money to a company like Comcast.

We've got Tachus in our neighborhood. $90/month for 1Gbps symmetrical fiber to the home. There is a $500 fee for running the fiber to the house, but its free if you stay for a min of 6 months (pay back prorated install cost if canceling early). Otherwise its a month to month deal, no contract, required modem is provided free of charge.

I'd like to switch, its half the price and more than twice the download speed, but they don't offer a business tier. I'm staying with Comcast Business, even though there is no SLA agreement, because call center wait times are much shorter, and you get much better techs on the line. I've only had one issue, and it was fixed within 30min. I can't risk the "regular isp" help line when my job requires consistent internet access.

Actually, does it require new law? I've yet to find someone talking on this topic, referencing copyright in corporate training. I'm involved in creating new technical training material a few times a year. Even though this is only for internal educational, I can't just use anything from the internet as it automatically* has copyright, and we are not an "accredited learning institution".

I'm not a lawyer, but how is a company training a 'system' legally different from training an employee? both are to increase company profit. In each case (human training vs AI training) copyrighted work was required for training, which violates copyright law for training an employee, unless you get the copyright holders permission. At our company, there is a library of images, icons, audio, etc. that we can use for any corporate purpose, as we own rights to use them.

I am honestly asking: how can you use an image to legally train the AI, when you are violating copyright if you use the same image to train a human?

*yes I know I'm simplifying e.g. not everything gets an automatic copyright, but since an artists work does, I'm skipping details like this.

"Anyway, when you put something online it becomes public domain, and thus fair game." No, it does not. Just because you post an image online or show it in a public space does not make it public domain.

"Most material found on the Internet is protected just like any other material (unless otherwise indicated). Text, charts, graphs, tables, photographs, music, movies, graphics, postings to news groups, blogs, e-mail messages, images, video clips, and computer software do not lose copyright protection simply because they are posted on the Internet." https://www.lib.sfu.ca/help/ac....

"When a photograph is in the public domain, it means that there is no copyright, and anyone may use, copy or distribute the photo for any purpose. There is nothing about posting photos online that inherently places them in the public domain." https://photocopyrightlaw.com/...

I don't see it as a negotiation ploy, just a regular business decision.

I've been involved as a subject matter expert with our R&D to determine the market potential of multiple products, platform support, upgrades, etc. If the estimate is that all of the work list requires 100k man hours, but the R&D team headcount and budget can only do 60k of man hours. You then have to prioritize what to do based upon existing customer support, road maps, technology releases, ROI, etc.

At my university, lab computers were suddenly not working all over campus. Both cheap desktops, and the far more powerful computers used in engineering labs for simulation. Someone or group went around and managed to take the RAM and CPU from them. The solution the school adopted, was to have the engineering school to make half a metal box (with appropriate slots for CD, USB, etc.) that got bolted to the desk over the top of the computer, and two opposing bolts had padlocks to try prevent their removal. Sure, its not infallible, but it was enough to make it require too much time to access.

While I don't dream about it, anytime a problem's solution eludes me, I give up till the next day. The next day, after a good nights sleep, I know what the solution is.

It may be in your mind now, but I assume you had legal access to the information so you could learn it. Its not that the LLM has information that came from copyrighted material, its they didn't have permission to use the material for commercial purposes in the first place.

At the least, the CEO "Marc Benioff: 63k Miles in 60 Days" https://www.radaratlas.com/mar...

Responds silently.... How about not reacting in the first place unless asked? I've got mine setup to know my voice, yet someone on my conference call says "firefox" and suddenly google has a lot to say. Certainly wasn't my voice. Certainly didn't say "Hey Google" either.

I seem to have a problem with getting google to do what I want at the best of times. "Hey google, play music ABC on XYZ speaker", googles response "Here's a video from YouTube...", playing on the Nest Max where it has as screen, not on XYZ.

At least it works for my limited scheduled tasks, and to turn on/off the rooms lights

It will succeed. This has been announced by GS1, the body responsible for the UPC standard for retail and for assigning codes to companies/products. They've mandated that products will no longer use UPC in 2027.

Who are "they"? GS1 the group responsible for the standard used for retail barcodes. If you have a new retail product, you have to buy a UPC code from them to ensure no duplication. https://www.gs1us.org/industri...

Agreed, no company should ever damage others property without legal backing. Every business I've seen has signs stating that if you block XZY, you'll be towed. Tow trucks are always there extremely fast.

I was in the middle of my education in the 90's when AI was popular again. I realized its no better than the 80's or even the 60's "AI", its just an improvement of hardware that's available. It's the same today, neural nets have been around forever, just not with this level of hardware. Today, it does look like the hardware is powerful enough for very good outcomes. I unfortunately don't think it is going away, with faster hardware, its just going to improve at least up to a point.