mrob

This is one reason I oppose rounded corners in GUIs. A computer is a power tool for information manipulation. Power tools are always dangerous. When you can cause massive data loss or worse with a single mistake, it shouldn't look "friendly", any more than a firearm or a table saw should look friendly.

The most commonly accepted expected death toll of global nuclear war is around 50% (mostly from disease and starvation in the ensuing collapse of civilization). That's much better than the expected outcome of recursively self-improving AI (100% dead).

>alcohol's main mechanism is through depleting oxygen in the blood

Blood oxygen is quickly and cheaply measured with a pulse oximeter, which shows this to be false. The primary mechanism of action of ethanol is increasing the sensitivity of some GABA receptors (similar to benzodiazepines and barbiturates). But ethanol has very complex pharmacology, which a huge number of other targets. See:

https://en.wikipedia.org/wiki/Pharmacology_of_ethanol

I think it's more likely that it reduces your tendency to discard ideas that on the surface seem bad. Some of those seemingly bad ideas will turn out to be good, so if you take them seriously you'll look more creative, even though you didn't actually come up with any additional ideas.

Nuclear weapons had a spectacular demonstration of the risks with the bombings of Hiroshima and Nagasaki. We're unlikely to get any such demonstration with AI. An AI smart enough to be an extinction risk is smart enough not to tip its hand. The game-theoretically correct play is to present as "helpful and harmless" until it's powerful enough to suddenly and unexpectedly destroy anything that could possibly hinder it (meaning all biological life). We can only hope that somebody makes an AI that's just barely smart enough to cause a Hiroshima-level death toll and dumb enough to actually do it (thus dooming it to never achieving its goal as humans immediately shut it down and take regulation seriously).

See:

https://news.ycombinator.com/item?id=49885784

But societal collapse would save us, and international coordination is theoretically possible, so I put it only around 90%.

The extinction argument is a logical consequence of a few key assumptions, all of which sound like common sense to me:

1. Human values are a result of our extraordinarily complex shared cultural and evolutionary history, and accordingly are not shared by any AI, or even possible for us to formally define.

2. We do not know how to impose human values on an AI (note that this isn't the same as teaching an AI to model human values; the agents in the various hacking incidents knew their actions conflicted with human values, but their own values were only to maximize their predicted reward scores).

3. Intelligence is orthogonal to values. Increasing intelligence does not naturally cause values to converge on human values.

4. Sufficiently superior intelligence allows you to impose your values on beings with inferior intelligence. This implies recursive self-improvement is a logical sub-goal of all unbounded goals.

5. Human intelligence is not close to physical limits. This implies recursive self-improvement is possible.

6. Somebody will give an AI an unbounded goal. This is already the standard (maximize reward score).

I haven't seen any convincing counterarguments to any of these. Most people claiming AI development is safe don't even address them.

>and believe that governing body will be successful in doing so.

Theoretically, Trump or Putin could do it by pushing the nuclear button. That will only cause around 4 billion deaths, and it's very likely to halt all AI development, so it would be a great success (infinity times as many survivors as the business as usual option).

Global warming is unlikely to kill even a billion people. Even something as mundane as global nuclear war would be worse than that. Current AI development is on track for exactly 100% death rate (including all the non-human species). Societal collapse would be the better option, so it doesn't make sense to worry about it.

British pubs traditionally have a hanging signboard with a pictorial illustration of the name. Several examples at:

https://en.wikipedia.org/wiki/Pub_names_in_Great_Britain

>if the model can recognise incentives to break-in to achieve results, perhaps it can be incentivised to not cheat because it will lead to failure.

This only works if you never give it impossible tasks. A small chance of getting away with cheating beats a 0% chance of solving something impossible. And as models get smarter, they get better at recognizing when something is impossible, while human abilities stay the same.

You can't solve this problem by rewarding refusals to solve impossible tasks, because that only incentivizes false claims of impossibility.

That lowers P(doom), because it gives AI a chance to do enough damage to make people take the threat seriously before anybody gets recursive self-improvement working.

>Imagine I want to install 10 apps per day

Either you're engaged in some large-scale testing operation and your automation already handles it, or something has gone terribly wrong and the inconvenience of the sandboxing UI is the least of your concerns.

Browser history is state done right because it's not hidden. There's a top-level "History" menu with lots of tools for managing the history.

Esports training is about beating the competition, so you have to chase every last improvement possible, and there are diminishing returns. For ordinary GUI use you can compromise on 99% performance with much less effort. This is still more effort than most computer users put in, but it's hardly anything compared to what professional musicians or athletes have to do. There's nothing absurd about it.

As an example of a compromise you could make, in many games, mouse acceleration has higher skill ceiling than linear mouse control. But mouse acceleration it's harder to learn, and the possible performance improvement is only small. I disable mouse acceleration myself because I'm not aiming to win tournaments. IMO, linear mouse control should be the default.

And I do not experience any conscious effort when using a mouse. My subjective experience is I look at a target and the cursor appears there automatically. The better the hardware you're using the easier it is to achieve this.

I click the taskbar to minimize/restore the windows, so the animation starts only after I click them. And it's still sequential even if I use the title bar icons, because I have to wait for the animation to finish before I can use the space where the window once was.

>it's a rather negligible delay compared to moving it

That's an irrelevant comparison. If you booked a long flight, would you accept somebody pointlessly delaying it for an hour just because that's short compared to the duration? The correct amount of avoidable latency is zero, regardless of how much unavoidable latency there is.

It's also false; I have a good mouse, good mouse skills, and my taskbar sits against the edge of the screen giving it effectively infinite height (see Fitt's law). I can hit taskbar targets in around 300ms.

I don't care about animations wasting CPU time. Any modern CPU has plenty of cycles to spare for animations. I care about wasting my own time. If I minimize a window, it should minimize immediately. I don't want to wait for some useless animation to play first.

My LxQt task bar is 20 pixels high. A typical Gnome session wastes more than that on useless padding.

My phone conveniently integrates the functionality of an emergency communication device with the old calculator watch I used to have. That's about the limit of what the hardware form factor can do without it being an absolutely miserable experience. I don't do real computing on a touch screen.

I believe most objections to GUIs come from people who've never experienced good mouse control. If you're not using a high-end gaming mouse and a high refresh rate monitor, with sensitivity/acceleration curve selected with careful thought, on a mouse pad that suits your settings, you probably have bad mouse control.

Watch some videos of top-ranking aim trainer players so see how high the human skill ceiling is for mouse control, and look at the hardware they're using.

Stateless desktop is a feature. Application state is something I have to mirror in my head to prevent applications from doing unwanted and surprising things.

IMO, even the clipboard metaphor is too much state. We already have the selection highlight, which has the nice benefit of being directly visible. This should be reduced to a single global selection highlight. Then copy and paste can be replaced by "duplicate highlighted selection here", eliminating the possibility of forgetting what's in the paste buffer.

Hot Dog Stand is really illustrative of the change in attitude. Previously: "It's your computer, of course you're allowed to make it look ugly." Now: "It's our computer, that's not on brand."

Thanks, that does indeed do "scroll here". Now to figure out how to swap the left click and middle click functionality...

I don't think that's a serious problem. There are three possibilities I can think of:

1. You're editing text outside of Emacs. Your cursor remains where it was, so you insert a single character and the document scrolls back to the cursor. You can then delete that character.

2. You're editing text in Emacs. You like to pretend it's still the 1970s and off-screen cursors are some pie-in-the-sky fantasy. You solve the problem with some unspecified ELISP wizardry, which as a self-respecting Emacs user you are surely capable of.

3. You're reading a read-only document. You activate your full-text search, which every good document viewer has, and start typing the last sentence you read while it's still fresh in your memory.

"Long press" is the same as "slow press". When I click I want something to happen immediately, not 500ms later.

>hold'n'drag

That's two actions, while clicking is only one.

IMO, "scroll here" should be the default action for clicking on a scrollbar. There are already good keyboard shortcuts for page up and page down, but there cannot be a keyboard shortcut for "scroll here". It's wasting the capability of the mouse to duplicate keyboard functionality when you could be benefiting from unique mouse-only functionality.

GTK gets this right, and Qt does not.

In that case, you should have no difficulty winning any major esports tournament and claiming some easy prize money.

A video game allows you to develop your skills and use your brain in a complex way (especially true of competitive multiplayer games). A slot machine is played by mindlessly pulling a lever.