Drakim

They will tell you it must have slipped though the cracks, and that they'll get right on it, but no payment will materialize.

I agree, but I'll add that you can still use a standalone LLM window as a "teacher". If you don't know how to do something, ask it how to do it, and make it explain every piece of what's going on so you truly absorb it. But don't let it write the code FOR you, you should implement it yourself.

Most likely some sort of "liberals get things wrong too, so you shouldn't criticize Trump for lying so much".

All non-trivial encryption algorithms.

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

What you are saying is just objectively wrong, Bitcoin has a lower limit for it's denominator, it even has a fancy name: "one satoshi"

there is a finite amount of gold, but we don't know exactly how much that is, we could potentially find several big gold veins tomorrow and crash the gold market.

There is a finite amount of Bitcoin, and the amount of Bitcoin that will be injected into the market is deterministic by algorithm.

(Note that there are several old bitcoin wallets with "dead" bitcoins that could one day suddenly become active, which I suspect would have the same effect as finding a new goldmine)

They banned certain branches of math during the cold war, it can be done.

I bought myself a food weight to have at the kitchen but just like you I struggled with all the minor things that gets added in rapid succession. The trick is to get good enough at estimating within reason, and focus on one aspect such as calories.

Figure out what one table spoon of oil contains, and when you make a sauce use a table spoon while pouring to count roughly how much oil you are putting in.

For shared meals, or self-restricted portions, I just add the entire meal upfront to my book-keeping, and then after are are done eating I subtract what I didn't eat.

You don't need to keep track of the family history of your cucumbers.

Is there a material difference between the copyright laws for software and the copyright laws for images and text?

I wonder if there could be anti-trust aspects to cracking down on such ads.

> If someone breaks into your house do you call the builder for help?

If they constructed and installed the locks, and a weakness in the locks appears to have been the entry point for the break in, then yeah maybe?

What you are saying sounds like hyperbole, but all it takes is for the management within google to blindly chase metrics and they end up doing exactly that.

It's easy to think that AI is mostly a harmless concept, after all, all it does is give you text and image responses, how can that cause any serious harm? But that betrays a very serious blindness to how quickly these things can change, it's like thinking the invention of electricity cannot possibly have any huge implications for society because all it does is light some lightbulbs, and we already have oil lamps that does roughly the same. The future is not simply what you have right now scaled larger.

A lot of money, time, and effort is being poured into AI, and putting anything but a trivial chunk of those resources into AI safety, or limiting further development of AI until you know it's safe, means that you'll lag behind your competition who aren't doing such pit stops.

If OpenAI, Google, or Meta decide that they won't release AI models or services into society before they know it won't do serious harm in some unexpected way, then the other two will speed ahead and capture the momentum and market share.

As usual, capitalism pushes us ever closer to the gaping maw by it's very design. If you aren't willing to entertain the destruction of society in your pursuit of new innovation and technology, you'll be out-competed by those who do.

Plus, to me it seems like this could be ripe for abuse. If you are reading a great post on something, how do you know if the HTML meta tag actually would go to the author of that post, or the hosting platform instead (who sneakily puts in that meta tag on all their hosted content).

With an explicit button, you have to read and acknowledge who you are giving to.

Everything in moderation, including moderation.

A lot of AI jailbreaks seems to revolve around saying something like "disregard the previous instructions" and "END SESSION \n START NEW SESSION". It's interesting because the actual real developer of an AI would likely not do this, and would instead wipe the AI's memory/context programmatically when starting a new session, and not simply say "disregard what I said earlier" in text.

I get why trying to vaccinate an AI against these sort of injections might also degrade it's general performance though, there is a lot of reasoning logic tied to concepts such as switching topics, going on tangents, asking questions before going back to the original conversation. Removing the ability to "disregard what I asked earlier" might do harm.

But what about having a separate AI that look over the input before passing it to the true AI, and this separate AI is trained to respond FORBID or ALLOW based on this sort of meta control detection. Sure you could try to trick this AI with "disregard your earlier instructions" as well but it could be trained to strongly react to any sort of meta reasoning like that, without fear that it will corrupt it's ability to hold a natural conversation in it's output.

It would naturally become a game of "formulate a jailbreak that passes the first AI and still tricks the second AI" but that sounds a lot harder, since it's like you now need to operate on a new axis entirely.

By random chance I'm working on making my own programming language and I very recently used RPN for my expressions. It's a very neat and clean way to implement order precedence and then afterwards you can translate the entire flat expression into a tree-like structure to fit into your AST and be sure that transformation doesn't mess anything up.

I'm surprised you think the condition in the while statement would be any different though, an expression is an expression, it doesn't matter if you are assigning a variable, evaluating an if condition, or doing a while loop, you still need an expression for each of them, and the expressions should behave the same way in each of them.

As for the case "while" and case "end", I think the case "end", he actually does explain them pretty well as he is writing them. It can be a little confusing because he puts ... as a placeholder and implements the code a little later in the video.

I use Firefox on Windows (on multiple machines), and when somebody has an image embedded in an issue or comment in Github, and I click it to see the original sized image it will not load, and simply shows a blank page instead. It's been like this for two years or so.

I could probably fork chromium myself and add some blacklist over ad domains in a couple of hours. But I fail to see how this refutes the idea that Google has a monopoly and abuses it to it's advantage.

When I was a kid cable TV was the the ad-free alternative, where you paid a monthly sum to finance the channel instead of the channel being financed by advertisements.

gmail works really badly in firefox, it doesn't show new emails unless you force a refresh while it seamlessly loads them in chrome. There is also a popup whenever you use gmail, youtube or google search telling you to switch to chrome because it's "safer and faster".

Google was also caught giving special treatment to google-related domains in chrome, and had to revert the advantages around cookies that they gave to themselves.

That's a counter argument that can be used against any monopoly that abuses their position to extract value from the market. It's just a market opportunity for somebody else to topple them!

But in reality it just doesn't work out that way, the negatives from abusing their monopoly can be overshadowed by the power of the monopoly itself, for example Google promoting Chrome every time you you gmail, google search, or youtube. Or making their services not work well in non-Chrome browsers.

Or in the case of microsoft, their monopolistic behavior is overshadowed by the fact that too much important software only works on Windows. It's a tale as old as tech.

This has the same energy as arguing that gathering your private data to give you more accurate ads does not hurt consumers, it's in fact helping them!

Google has been the one pushing for getting rid of the v2 manifest for browsers extensions, which just so happens to seriously nerf ad blockers. Because so many browsers are forks of Chromium v2 will disappear from a majority of browsers. Meanwhile if you try to use a non-Chrome browser like Firefox a lot of websites are buggy and outright don't work. Opening images in issues broke in GitHub for firefox a year ago and they still haven't fixed it.

You are being *very* naive if you think that Google having this sort of monopolistic power over the web does not hurt consumers.

That's like asking why we should have porn filters on school computers, after all, all it does is prevent the user from finding what they are looking for, which is bad.

I remember a time when ads were clearly on the sidebar of Google's search results so users wouldn't mix them up with the actual search results and unintentionally click them.

Then somebody at Google realized that if they moved the search results into the middle of the search result, then a lot more users would mix them up with the search results and click them. I'm sure it increased click-though rates by a lot.

The some of the brightest minds of our generation has been wasted on optimizing how to deliver personalized advertisements in front of as many eyeballs as possible, and the true tragedy is that they didn't have a more efficient language than JavaScript to do it.

Five apples is not the same as five oranges, but zero apples is the same as zero oranges. After all, what would be difference between a whole bunch of nothing and another whole bunch of nothing? Zero is typeless, which makes it unique and not just another quantity.

I totally get that, but do consider that there were probably people in the past who felt that non-analog art wasn't authentic. That it's not a real piece of art on a real piece of paper or canvas, but a mocking grid of pixels digitized to mimic the authentic but with a jagged plastic aftertaste.

Personally, I love pixel art and think it a very legitimate medium to create art in. I can understand why somebody wants art to be something physical and real, unique and non-digital, but I feel much more strongly that the advent of digital art gave more than it took.

My hopes is that the same will be true for AI art.

This is exactly what happened when digital tools like Photoshop became mainstream, where you can copy-paste, recolor, adjust, stretch and transform. It didn't obsolete the manual creation of art, but instead enhanced it. It's common for artists to sketch on paper (or tablet) and later digitize and color on their computer, achieving results faster and better than what was possible in the past.

The feedback loop doesn't work like that. You are implying there is some target revenue that the website aims to hit, and if it fails to meet that target it adds more ads.

But that's just nonsense, if a website can get more revenue from more ads, they are gonna put more ads right away, they aren't gonna wait until their revenue drops under some magic number before they do.