I might or might not disagree with you, but I have nothing to go on.

I think there’s a distinction to be made between solutions which are client-side, versus those that are server-side, or at least have a server-side component that is functionally necessary.

For the first category, the termination of a subscription shouldn’t lead to denial of access to the current build of the software: it should remain functional in the absence of lower-level updates. Ideally this software is sold as a one-time purchase, having life-time free updates bundled, or sell upgrades on a versioned basis; but a subscription fee for future updates seems perfectly reasonable too, if the previously stated conditions are met.

For the secondary category I believe it’s in the service’s right, to provide it’s products on the basis of its own terms and conditions; that is, if it doesn’t employ morally questionable practices, like: outpricing other services using venture capital, for the purposes of attracting a user-base for future exploitation (by raising fees, or by turning the user into the product; despite an ever decreasing value proposition: i.e. venture capital enshittification).

Algorithmic feeds are the problem here, which are exacerbated by features like autoplay and infinite scrolling. Users should by default only be served content from channels they explicitly chose to subscribe to. Algorithmic feeds should be separate from this subscription-based feed, and only bleed over if subscribed to said channels; which shouldn’t be any different for shorts (only showing shorts from subscribed-to channels, if not hidden altogether when preferred). Simply put: make timelines finite again, by not endlessly resupplying it with spam; which includes the unhealthy amount of “sponsored content”.

In the context of spy satellites, (nuclear) weapons and drone bases in orbit, I honestly can’t wait for the Kessler effect to materialize. Russia has demonstrated it could destroy a defunct Soviet satellite in orbit using a missile, and with tensions rising, alongside the number of objects placed in orbit, it doesn’t seem like a matter of if, but rather a matter of when it unfolds.

What is this, 1984’s Physical Jerks?

Most “name brands” have long been acquired by large umbrella corporations, and shortly after doing so, the “brand recognition” is often leveraged to market white label products; which is increasingly the only differentiator between it and off-brand products. That, and the price-difference: simply paying more for a meaningless “name brand”, on an equally inexcusably poor quality product; besides a slightly less shitty customer experience, hopefully.

I really think it’s poor design to purely filter on appearance of brands, rather than actual brand reputation. Yes, it might serve as an overgeneralized indicator for questionable reputation, but marketable brands shouldn’t be treated as reputable either.

Excellent video, and very applicable indeed! :)

But it naturally skews the results, especially when baby boomers and gen X alone constitute 57% of all respondents: which are typically less digitally literate, and therefore even less considerate of downstream consequences. Am I missing something? Because I cannot find any reference to another study, if that is what you mean. I’ve seen my share of pseudo studies that governments desperately cling onto, and by all means, this appears to be another one of them.

The statistics appear to originate from from this pseudo study. All of the questions related to supposed “protection of children”, are framed within this context, without concretely considering the possible downsides; which strongly biases participants’ views. I also like how it typically reduces ages 18-54 to a single bracket (while statistics on page 18, demonstrating a trend of younger respondents, responding with reduced strength), and especially how it doesn’t consider minors to be Canadians (18- being asked absolutely nothing, at all…).

I despise Alex Karp and Palantir, but this feels like a pathetic attempt to discredit him, for reasonably throwing shade on grossly over-valued AI products, unsuitable within the context of critical infrastructure and enterprise. His Berkeley faculty member argument simply seems to be an analogy, for other models being capable of performing the task, but it simply not going to happen in his belief. This article appears to serve stakeholders of the AI bubble.