HN user
People go on hikes in areas where there are multiple days between charging opportunities.
Hiking often also occurs in areas with bad or no coverage causing higher battery consumption with the phone trying to connect. If you don't mitigate this (with eg. turning off or enabling airplane mode) this will burn through battery much faster than the usual city dwelling.
Apple was also in the list, so still relevant.
Cant wait for Bendgate 2025 edition
Here is an English translation of the Norwegian law around this topic: https://lovdata.no/dokument/NLE/lov/2020-05-29-59 (Chapter 4 is of particular interest for this discussion)
I assumed the US had something similar, but seems I was wrong.
I should have said liability, it would have been more precise, but I'd argue that liability is a form of responsibility so I don't think your correction is warranted.
We do put additional editorial standards on news publications. This puts legal responsibility for the published content on the publisher.
It doesn't seem like that big a step to apply a similar standard to advertising platforms. Advertisers have failed to selfregulate the ads they choose to publish and it is infeasible to use the court system to judicate every false ad (that would be millions of court cases). Ergo you do the obvious which is to make the advertiser name a human editor who holds legal responsibility for published ads (on behalf of the company).
Now you can sue the advertising company (eg. Google) for millions of false advertisements at once.
Stealing a physical thing = the previous owner can't use it.
Copying a thing or accessing a platform = the previous owner can still use or sell it.
Even if you consider it unethical access, the comparison to stealing really misses the mark.
It requires less political capital to repurpose an existing system than to introduce a new system for a specific purpose. See for instance the number of times Chat Control has failed to become law.
Reverse chronological + subscription (ie. the user must actively make a choice to follow some channel or creator to get them in their feed). This is how most platforms started, and while there were still issues (eg. rewarding frequent posting) they seemed a lot less problematic than what we have today.
The main issue isn't the misinformation or disinformation; it is how quickly you can amplify reach and reach millions. Reverse chronological + follows based on active user choice would largely address that issue.
Say you want to quit sugar or smoking. Would you still buy cigarettes or chocolate and carry it around in your pocket everywhere you go because you should rely on willpower to beat the addiction? Very few people do that because you become vulnerable when your willpower is at it's weakest.
Usually it works better to exercise willpower to constrain your future self's available actions. For example, by not buying chocolate or cigarettes when you are at the store.
The same principle applies to your phone. Use your willpower to constrain what your future self can do with it.
Wasting peoples' time and attention for not using one (out of personal choice or necessity) feels like overreach.
This already happens with every ad successfully shown to a person. Why don't you criticize the ad business for much more extensive overreach instead of someone doing harmless activism on their own website?
They have invested huge amounts in the AI bubble which they have to justify to shareholders and / or investors. The easiest way to do this is to integrate it into existing products which already exists so they can claim millions of daily active users.
Same reason Google has added AI summary to their most used product: search.
Who is to say they don't? Most of those areas have a less developed health infrastructure, so I'd expect the statistics on dementia diagnoses to systemically underreport the issue.
The average age in developing countries is also lower (due to both lower life expectancy and higher birth rate) which is a further confounding value to skew the numbers. Dementia is an old people game, even if it is true that air pollution worsens it.
Git is designed so that you always have the full code you're working on copied to your local machine. Github being down for a short time from time to time should be only a minor inconvenience.
Sunc cost fallacy. If they admit that people don't want it (at least enough to cover the costs of it) share holders will question all the investments into Copilot. So instead they push it down our throats.
As long as they haven't been bullied into the corporate equivalent of suicide by the "justice" system it's not disproportionate considering what happened to Aaron Schwartz.
If anything it's too little based on precedent.
Multiple uses of em-dash. Must be written by AI.
Who passes those laws? The politicians bought and paid for by the same companies?
One issue with capitalism — and every alternative system would have its own variant of this problem — is rent seeking. Players become so powerful within the economic system that they no longer play by the rules, but rather write the rules to their benefit. This allows them to extract value instead of creating value.
This is a run-away positive feedback loop which is kept in check by regular revolt and redistribution of wealth (see eg. the labor movement).
Paper?
Even if pw and 2fa secret are stored together you get better protection against phishing because you never enter the full 2fa secret into a website.
We've literally had convenient online banking for two decades at this point without any DRM.
Don't rewrite history.
Publishing publicly doesn't surrender copyright…
Wikipedia continued existence is not dependent on ads or clicks.
If AI kills the ad-funded business model of the internet, it will be one of the few good outcomes from AI.
Good riddance.
As a consequence of the freedom of driving cars we got seat belts. But we did not prevent everyone from driving.
Even the self-proclaimed bastion of libertarianism, the US, has laws against booby trapping, so that is obviously not true.
That's not what world model means, neither in psychology or artificial intelligence — two fields Scott writes a lot about so he should know how the term is used or define how he uses it if he uses a non-typical definition.
We need that user data is only usable within the ToS under which they were originally collected. No more unilateral altering of the agreement after the fact.
Collecting user data should be a liability, not a enormously profitable endeavor.