> Cool man, but it might surprise you to find out that people knowingly opt into receiving this stuff and actually consent to it.
Consent is more than pressing 'Allow' on a notification pop-up. It's often not even informed consent, as those pop-ups are usually a part of some onboarding flow where users are just trying to get to the value the app promises and pressing 'ok' to everything.
Even if people do indeed want notifications at the time of the ask, one doesn't really know if the message provider will wind up spamming, that's a matter of trust. And once opted-in, even if the users no longer want notifications, a lot just don't know how to turn them off. People are often incredibly accepting of sub-par experiences like this because of the friction and capability demanded of them to opt-out. My parents get tons of spam notifications that would pass your test of 'knowingly opt into receiving' but that when asked they say they do not want.
Finally there's myriad dark patterns that tons of apps use, like changing and resetting notification preferences among others.
I'd hazard a guess that observed opt-in rates far exceed users actual desires, so I wouldn't put much stock in them. I do agree that there are some people that like them tho!
Fwiw I've worked on both the delivery side (OneSignal) and developer side (Margins) so I've lived these choices and trade-offs. My believe is in terms of power dynamics, senders generally don't deserve their power to interrupt and should not possess that power. At best, they offer opportunities, which ideally are verified somehow prior to being presented to users. I'm happy that devices and ecosystems are moving in the direction of triaging and filtering sender content, as power needs to lie in the user's holistic, most pre-frontal cortex driven expression of their desired experience, and not just one moment's opt-in button they pressed.
Thank you for writing the article, good discussion points.
Yeah, that's true about the allow, and for sure marketing and product teams are deploying misleading consent priming which doesn't fully explain to the user what they're actually allowing in the first place, or setting baselines that are too permissive vs what the user is expecting.
> I'm happy that devices and ecosystems are moving in the direction of triaging and filtering sender content, as power needs to lie in the user's holistic
I don't disagree necessarily, but I see it as them putting themselves in a position to act as a toll collector, which has already happened with email and web search and is only getting worse with the introduction of LLM's into both of those things.
It's a bummer this article ended up doing much better than my email one, as I think that might better position the problem in a lot of user's minds and highlight just how much surveillance is sitting on top of those free inboxes.
The generations system in III was amazing. I didn't care that the art wasn't top notch, I loved the game and seeing the story over three generations. Decades later I wound up making a storytelling game mostly based on it.
My favorite thing about 3 is that the soundtrack has tracks based on the number of characters in your party. Gain a character? Suddenly the music is richer. Lose someone? The music sounds more lonely.
Bringing your own cup is also healthier, because paper cup linings are made of plastic that when exposed to hot liquids are likely to turn into microplastics.
My rule is: 1) keep hot stuff away from plastic, and 2) assume everything is/has plastic unless you know for certain.
I did this calculation a bit ago and don't think frontier models are just a few MacBook Pro generations away. Yes numbers reliably go up in tech in general but in specific semiconductors & standards have long lead-times and published roadmaps, so we can have high confidence in what we're getting even in 3-4 years in terms of both transistor density and RAM speeds.
In mid-2028 we have N2E/N2P with around 15% greater transistor density than today's N3P, and by EOY2028 we'll likely have A14 with about 35-40% density improvement.
Meanwhile, we'll be on LPDDR6 by that point, which takes M-series Pros from 307GB/s -> ~400GB/s, and Max's from 614GB/s -> ~800GB/s.
Model improvements obviously will help out, but on the raw hardware front these aren't in the ballpark for frontier model numbers. An H100 has 3TB/s memory bandwidth, fwiw
What do you need 3 TB/s memory bandwidth for in a single user context? DeepSeek V4 pro (the latest near-SOTA model) has about 25 GB worth of active parameters (it uses a FP4 format for most layers) which gives 12 tok/s on a 307 GB/s platform as the current memory bandwidth bottleneck, maybe a bit less than that if you consider KV cache reads. That's not quite great but it's not terrible either for a pro quality model. Of course that totally ignores RAM limits which are the real issue at present: limited RAM forces you to fetch at least some fraction of params from storage, which while relatively fast is nowhere near as fast as RAM so your real tok/s are far lower (about 2 for a broadly similar model on a top-end M5 Pro laptop).
At least this gives kids the chance to be kids and know a life without it before they encounter it.
Many of us grew up in the offline-is-default time, but our cohort will age out. Then we’ll only be left with people who grew up with these technologies shaping their lives and perspectives, who have little sense for the alternative.
The window of time is closing for us in this cohort to use our understanding of what life was like without these technologies to advocate for a healthier environment for kids.
Please. At best HN has a very small subset of the problems in social media, and its positives easily outweigh its negatives. This is a well moderated forum with a lot of bright people and industry experts that a young person could learn from by observing conversation and debate. It bears a great deal of resemblance to historic methods of learning by watching experts interact and debate. Tons of pedagogical value is here for a young person to latch onto.
A most obvious difference besides that is HN isn’t a nonstop feed of short form video appealing to the insecurities of teenagers, using notifications and social feedback loops and the suggestion that you’d be missing out on what your friends are up to if you left.
HN doesn’t even let you follow people and barely lets you know who they are. It’s centered on ideas, not people. HN and social media are almost nothing alike.
Totally. Today’s social media is not the same as last years etc. Read Meta’s quarterly reports and they brag about Reels increasing time spent on site by 30% in a year. That’s not even considering the other ill effects like giving kids a firehose of all the worlds problems when they’re not yet equipped to handle that information, which causes them to internalize those things, making them feel like things are fucked, that they’re responsible, etc. It’s psychologically devastating. And so many other things! Let kids be kids.
Consent is more than pressing 'Allow' on a notification pop-up. It's often not even informed consent, as those pop-ups are usually a part of some onboarding flow where users are just trying to get to the value the app promises and pressing 'ok' to everything.
Even if people do indeed want notifications at the time of the ask, one doesn't really know if the message provider will wind up spamming, that's a matter of trust. And once opted-in, even if the users no longer want notifications, a lot just don't know how to turn them off. People are often incredibly accepting of sub-par experiences like this because of the friction and capability demanded of them to opt-out. My parents get tons of spam notifications that would pass your test of 'knowingly opt into receiving' but that when asked they say they do not want.
Finally there's myriad dark patterns that tons of apps use, like changing and resetting notification preferences among others.
I'd hazard a guess that observed opt-in rates far exceed users actual desires, so I wouldn't put much stock in them. I do agree that there are some people that like them tho!
Fwiw I've worked on both the delivery side (OneSignal) and developer side (Margins) so I've lived these choices and trade-offs. My believe is in terms of power dynamics, senders generally don't deserve their power to interrupt and should not possess that power. At best, they offer opportunities, which ideally are verified somehow prior to being presented to users. I'm happy that devices and ecosystems are moving in the direction of triaging and filtering sender content, as power needs to lie in the user's holistic, most pre-frontal cortex driven expression of their desired experience, and not just one moment's opt-in button they pressed.
Thank you for writing the article, good discussion points.
reply