Probably the best idea I guess as long as you can set the TV up without Internet.

The TV I currently have is Android OS but the built in Chromecast is noticeably lower quality. Not sure if it’s an older version or what.

Regardless, IMO the displays themselves outlast their software support, and I prefer to just plug in whatever the latest device.

I’ll also mention Android OS on my TV takes a full minute to “boot” and that itself makes me want to yeet it out the window.

My TV is probably going to kick the bucket in a year or two at most. Filtering “non smart TVs” on a site like BestBuy shows only commercial display options at this point.

Are there any well maintained projects out there that are able to replace the firmware on newer smart TVs to get rid of these features? I really just want a dumb display with an input for a Chromecast with CEC support (or similar device if Google decides to enshittify that platform with screensaver ads too).

This is giving me 1998 MS Publisher vibes and I’m here for it.

I remember when I was growing up

You can basically stop right there. You were young and naive, viewing the world through the rose colored glasses of youth.

The context is not the same. A snippet is incomplete and often lacking important details. It’s minimally tailored to your query unlike a response generated by an LLM. The obvious extension to this is conversational search, where clarification and additional detail still doesn’t require you to click on any sources; you simply ask follow up questions.

With Gemini?

Yes. How do you think the Gemini model understands language in the first place?

If it’s just that and links to the source, I think it’s OK.

No one will click on the source, which means the only visitor to your site is Googlebot.

What would be absolutely unacceptable is to use the web in general as training data for text and image generation.

This has already happened and continues to happen.

What then will they use to train it?

Good. Please proceed as quickly as possible.

How is this different from the capabilities of Tesla’s FSD, which is considered level 2? It seems like Mercedes just decided they’ll take on liability to classify an equivalent level 2 system as level 3.

It’s not enough.

I’ve never clicked an ad on purpose. I use DNS to block all the common click thru domains for ads.

This move by Microsoft will undoubtedly result in more Windows PCs infected by malware as people find tools to remove the ads and some of those tools will turn out to be malware.

If you don’t have prime you might not buy things from Amazon which would probably be a net loss compared to the potential ad revenue increase.

You can also mark designs as ready-only and they no longer count, so this limit is really 10 concurrently editable designs. I just keep everything read-only unless I’m actively working on it.

Although, if they do stop some mass shootings we’d likely never hear about it. So it’s not really that straightforward to understand what, if any, impact it has had.

I also made the pistol from Galactic Armory which is not in the above picture. Assembled from PETG parts and painted using Rubb n Buff metallic paints.

Thank you! This was my first large project like this, and there were definitely times when I got sidetracked too.

I scaled the helmet at 98.85% so that it would fit on the bed at 45⁰. The fit works for me with padding that is about 12mm thick. 100% scale probably would have been ideal to provide a little more airflow, but I wanted to avoid glueing and filling seams.

Parts were printed on Bambu X1C with Hatchbox Black PETG and using PLA for support interfaces. Typical 0.2mm layer heights with adaptive layers down to 0.12 to reduce banding on the top of the helmet. Scaled the helmet to the 256mm³ print volume as a single part and it fits my big head with enough room for interior padding. Models are from Galactic Armory.

Some actual print times for reference:

• Helmet: 30.1hrs
• Chest Plate: 16.6hrs
• Forearm: 14.8hrs
• Shoulder: 7hrs

Preparing for painting I used progressively finer grit sandpaper and wet sanding from 140-400 grit, alternating with Rustoleum filler primer. Finished with black primer and 800 grit before applying paint.

For paints I used Montana Gold Shock Black for a base layer, then a layer of Spray Max 2K clear coat, and I tried Montana Chrome Effect over this. It was too silvery, so I switched to a powdered graphite rub, alternating 2-3 layers of graphite rubbed into a semi-soft clear coat layer (after only 15 min of dry time per layer).

Leather parts and base suit purchased from UMBLeather on Etsy. Armor is attached with velcro.

The New York Police Department has reportedly seen a 60% increase in ghost guns seized from city streets for the past two consecutive years. NYPD recently traced some online ghost gun sales to a “ghost gun printing operation” filled with 3D printers and firearms nestled within a daycare center.

I have no idea from this if the increase in ghost guns was attributable to 3D printing, or if the increase even represents a significant number.

It also seems like this would be really difficult to enforce. You can buy hardware or kits to build your own printer, or you can buy an old printer off someone else.

All that said, if it represents a reasonable concern and they figure out a way where it’s not trivially circumvented (both of which seem unlikely), I really couldn’t care less about a background check.

I’m not sure how this is different from Google pushing popups in every Google app to switch to Chrome.

I’m sure there is a reason, but if you have a collection of mp3s, why not store them on your phone or stream them from the cloud? Why use an iPod at all?

Getting information into and out of those domains benefits from better language models. Suppose you have an excellent model for solving math problems. It’s not very useful if it rarely correctly understands the problem you’re trying to solve, or cannot explain the solution to you in a meaningful way.

A similar way in which language models are already used today, is to use their predictive capabilities to infer from your question which model(s) might be useful in responding, gather additional relevant information, and to repackage this information as suitable inputs to more specialized models or external systems.

This finally pushed me to stop supporting the platform with a premium sub and now here I am, checking out lemmy!