Marvin

JP234 said:

"GPT-4 also lacks knowledge of events after September 2021 and can't learn from its experiences. However, OpenAI has improved many of GPT-4's safety properties."

So, if it can't learn, and it's around a year and a half out of date, why is it better than a browser? Because it can write a term paper?

Humans don't get hired for jobs based on how good they are at using search engines but based on their ability to write term papers.

https://www.pcmag.com/news/chatgpt-passes-google-coding-interview-for-level-3-engineer-with-183k-salary

Search engines respond to requests with references/links/associations. AI models are knowledge engines that respond with direct responses to inputs.

There's much less of a need for a knowledge engine to be up-to-date because a lot of knowledge doesn't change over time like Mathematics, software development, scientific discoveries, laws of physics.

A few people are skeptical of AI just now because it's new and with new things, people immediately point to flaws and limitations and it gives an excuse to dismiss it. These models are in their infancy and are already able to do amazing things. They are currently consuming every piece of data that exists in the world.

It's fair to say that all they are doing is mirroring the knowledge, skill and creativity of humans and they will only ever be as good as what they are mirroring but when that's combined with a lot of processing power, it can recreate all of what humans do in seconds instead of days/weeks/years and that's good enough to be extremely useful.

twolf2919 said:

Why is it so hard for authors to get AR vs VR correct?  All the rumors about an imminent new Apple device have been for a VR or MR/VR product, not an AR device.  Heck, even the stock photo used for this article shows goggles with opaque front - something you could not have with AR glasses/goggles.

There's passthrough AR that uses front-facing cameras, which captures real-time stereo video at a high FPS and merges digital content into the feed. This is how AR currently works on iPhone/iPad. This can give better quality AR because it allows better matching between the digital and real images and fully opaque digital content.

The downside is that you are always looking at a display which can cause eye fatigue more easily than looking at natural light and the real video feed doesn't look as natural as looking through lenses.

Here's passthrough AR on Meta/Oculus Quest VR headset:

https://www.youtube.com/shorts/jUIE2l_9ig8

marktime said:

mknelson said:

marktime said:

Not really. Microsoft doesn't sell Windows for ARM. The only version that I am aware of is the Windows 11 Insider Preview ARM64. That's not a generally available version. There's a lot less to this announcement than is implied.

Parallels already had a simple install for Windows 11 Insider Preview ARM64 - super slick and super quick!

I think the main takeaway is Microsoft won't go out of their way to break ARM64 installs if they aren't on  computers with Qualcomm CPUs.

Agreed, the Parallels install and support for Windows 11 Insider Preview works very well. The integration between the two operating systems works extraordinarily well. What I object to is the first sentence of the article.

"You can get Windows onto an Apple Silicon MacBook Air." 

No you can't unless you are enrolled in the Microsoft's Windows Insider program.

That's what's changing:

https://www.parallels.com/windows-11-arm-apple-m-series/
https://arstechnica.com/gadgets/2023/02/microsoft-officially-blesses-parallels-as-a-way-to-run-windows-on-m1-m2-macs/

Parallels has an easy installer for Windows and it can be licensed through the normal Windows 11 site:

https://www.parallels.com/products/desktop/welcome-trial/

They support Windows 11 Pro and Enterprise.

Xavalon said:

Apple has over a billion customers. No way that 50% owns a Mac. 

Apple has even said as much:

https://appleinsider.com/articles/18/10/30/apple-passes-100m-active-mac-milestone-thanks-to-high-numbers-of-new-users
https://www.theverge.com/2021/1/27/22253162/iphone-users-total-number-billion-apple-tim-cook-q1-2021

The surveys normally target users in the US where Mac ownership is higher but it's very flawed to make it a general statement, especially when Apple puts out active user numbers now and then. Kids are likely to have iPhones and not Macs and won't be surveyed.

Apple currently has 2 billion active devices but this includes everything like Apple Watch, Apple TV.

If we assume 50% new users in each segment, iPhone should be around 1.3 billion, Macs around 175 million, iPads ship roughly 2x Mac so around 300 million. This leaves 200 million for the others.

The figure of 90% Apple customers owning an iPhone is viable because it's likely a lot will own some of the other products but if 1.3b = 90%, Mac ownership is closer to 15%, iPad 30%. Apple Watch is estimated between Mac and iPad.

amongthecows said:

This deal is fascinating. Phil Spencer has indicated it’s “…always been about mobile and casual." But realistically it is still about the 800lb gorilla that is COD. Yet, nothing placates Sony and they continue to lobby to kill the deal instead of invest in their own products and services. At the same time EA has made it clear that mobile is a waste of resources by shuttering two tent pole IPs (Apex & Battlefield) in the mobile space. So, what if Activision just holds an IP firesale? MS buys COD, Candy Crush etc and leaves the chaff behind. Can Sony and its lobbing machine prevent that too? 

It's strange that a company can be built with such important products in a market they aren't allowed to sell it. But the company is basically comprised of 3 heavyweight products/franchises:

https://investor.activision.com/static-files/7141e015-b629-47c0-a755-dc837cb54cbd

"in 2021, revenues associated with the Call of Duty, Candy Crush, and Warcraft franchises, collectively, accounted for approximately 82% of our net revenues—and a significantly higher percentage of our operating income"

Revenue is split roughly $2b yearly for each part - Activision (Call of Duty), Blizzard (World of Warcraft), King (Candy Crush). They also have other high profile franchises Overwatch, Crash Bandicoot, Diablo, Guitar Hero, Tony Hawk.

Call of Duty games sell 10-30 million copies. Candy Crush has hundreds of millions of players. World of Warcraft has a few million subscribers that have been playing for years.

Losing these games on a platform would easily lose 1/4 of the player base.

The same is true of Sony's exclusives though. They have exclusive franchises like God of War, Gran Turismo, Horizon Zero Dawn, Ratchet and Clank, Jak and Daxter, The Last of Us, Uncharted.

I would definitely say that Microsoft owning Activision Blizzard would heavily weight the good franchises on their side.

These kind of issues are going to keep appearing the more that platform providers and content providers consolidate, much like in the streaming video market. Platforms owning content makes them anti-competitive because they restrict access to content.

The regulator's recommendation is to split parts of the company off with the main recommendation that Microsoft can't buy Call of Duty in any circumstance:

https://www.gov.uk/government/news/microsoft-activision-deal-could-harm-uk-gamers
https://assets.publishing.service.gov.uk/media/63e376bdd3bf7f173ad1cee4/Notice_of_possible_remedies.pdf (note 14)

If Microsoft won't go through with the purchase without Call of Duty, it's clear they want that franchise and would likely make some or all of it exclusive to their platform at some point. Another alternative is that they agree to publishing Call of Duty on rival platforms indefinitely, not just 10 years as they proposed, in its entirety including all main DLCs.