Anyone else know what’s up with including Framework at the end of the post? That seemed to come out of left field. I checked and nowhere else in the article is it mentioned, so it’s not clear at all what his beef with Framework is.

Exactly. The concept of an LLM is perfectly fine, we’ve had its precursors like markov chains and neural nets for decades, the issue is how their being created and used these days, particularly the scale of it. LLMs are the shiny new toy the way blockchains were previously, so all the tech bros and utterly clueless CEOs are rushing out to embrace the hot new thing and demand it be shoved into every nook and crany because they think if they don’t they might not make every last cent they possibly might.

Quora was always that site where you’d search for something on Google, see it show up near the top of the results and you’d just instantly skip over it because you already knew whatever it said was a waste of time unless you were looking for a laugh. Half the time the answers would literally be people messing with the person asking the question, and the other half of the time the person asking the question would be the one joking around.

While that is true this news doesn’t actually have anything to do with gaming. This was Nvidia telling other corporations that they’re going to price gouge an extra 15% for server GPUs (ironically making cloud gaming potentially that much more expensive).

It’s the exact same problem you run into on any open or semi-open market but the outcome for anything with a good user review system will be the same. Look at Steam where asset flips and low quality clones flooded the store but 99% of them barely get a handful of sales if any. Instead all the money goes towards the games with actual fun content. People recognize low quality and low effort slop and AI reeks of it.

The only thing flooding markets with slop does is it makes it harder for the small gems to get discovered. Some of them still will, they happen to catch the eye of just the right person and spread through word of mouth, and the big companies will just shovel heaps of money into advertising, but it makes life miserable for the little guy trying to make a name for themselves.

I mean it kind of already does this accidentally, just look for em dash and/or emojis sprinkled all over the document and you can be pretty certain that it was AI generated.

The other thing I’m seeing here is that this will only be effective for large chunks of text, if you’re dealing with small snippets interspersed with human generated content there will be enough statistical noise to make classification hard without introducing a bunch of false positives and negatives.

It would make a good second (or third) layer of defense. The problem with it is that it’s reactive not proactive, so you need to download the ad first before it can decide to block it or not. URL blacklists are the superior option, but hard to maintain. The ideal setup would be a URL blacklist that’s automatically updated on the fly by an AI classifier so anything that slips through the blacklist gets ban hammered by the AI and added to the blacklist to prevent future wasted bandwidth. If the blacklist updates were shared the same way some of the ad blockers do currently it could be incredibly effective as it would be essentially distributed realtime ad blocking.

Eh, maybe. First for existing content it makes no difference, the creator already got paid to read the ad so whether you skip it or not doesn’t change the money that’s already in their account. The argument would be if people are regularly skipping the sponsored ads then companies might stop offering sponsorships. Considering this is currently limited to premium customers which a) tend to be the minority, b) still need to manually trigger a fast forward to skip the ad, and c) would likely fast forward through it anyway because they’ve already shown they’re willing to pay money to avoid ads, I don’t think this will actually make much difference. Worst case scenario it drives down the amount being offered for sponsorships a little.

The content creator won’t care but the sponsors might. YouTube has a similar system that skips over “commonly skipped segments” when you’re fast forwarding. In practice it just means you skip past the section of the video where whoever it is reads the sponsors ad copy. It would have gotten skipped anyway, this feature just saves you a few seconds of putzing around trying to find the exact spot where the ad ends.

So the company made their game marginally less shit which caused their revenue to dip very slightly and the stock markets reaction was “SELL!!!1!!!ONE!!! THEY’RE GOING UNDER!!!”. This is why we can’t have nice things.

Setting aside the fact this was definitely not unprompted, social engineering might be the one thing AI is actually better at than a human. That is after all one of the biggest uses for bots for a while now by way of spam bots trying to social engineer people into various activities.

Unfortunately there are very few actual CGM brands available. There are a bunch of different CGMs, but most of them are just relabels of other units, almost all of which are made by either Dexcom or Abbott. Even more unfortunately in my experience they wildly exaggerate the accuracy and lifetime of the units. Most claim an accuracy of +/- 20% which is pathetic for 14 days, but you’re lucky if it maintains even that terrible level of accuracy for half that time.

In my case I’m horrendously allergic to the adhesive that Dexcom uses on their units, so I can’t even use half the CGMs on the market, and I gave up on Abbott after their units started to struggle to even make it 48 hours without being off by more than 50 mg/dl.

No, you have failed to understand how either license works. Companies are forced to contribute back in the case of the GPL and the project can choose to accept that or not. In the case of the BSD license companies can choose whether to contribute back and if they do, then the project has the option to accept it or not. The GPL makes it always the choice of the project, while the BSD makes it always the choice of the company.

Basically his argument is that because some companies choose to contribute back to BSD projects even though they aren’t required to, and because some GPL projects choose not to merge some changes from downstream forks, the GPL and BSD licenses are effectively the same. I think the flaw in this logic should be pretty apparent. I think the GPLv2 vs. GPLv3 debate is a lot more interesting, particularly since Linus has come out against GPLv3.

The point isn’t to make it forensically undetectable, but to prevent the room temperature IQ cop from noticing what happened. If they “unlock” the phone and don’t see anything that looks out of the ordinary they won’t make a big deal out of it and there’s a chance they’ll let you go.

Don’t forget power it as well. Watched a video where someone did the math and to generate the power required of a typical data center you’d need a solar array something like 40 times the size of the one on the ISS. It would be an absolute behemoth of a satellite and that’s without taking into account the cooling which would likely double the size again. Of course once you get that big then you have to worry about keeping the thing in orbit because moving that much mass isn’t exactly easy. No matter how you slice it it’s a stupid ass idea that doesn’t work.

I had never heard of them but apparently it’s an “open source” AI platform. Basically AWS but aimed specifically at running LLMs.

Well first you need to invent AGI which is probably at least a decade away still, but once you do that then you negotiate with that AGI to carve out a little corner of the Internet and lock all of us inside it.

Claiming your AI product sucks and doesn’t do what it’s asked to do (aka hallucinates) is a bold move, lets see how it works for them Cotton.

If you’re influencing the outcome via some private mechanism that’s a crime and is in fact illegal (a little thing known as insider trading or cheating depending on how you go about it). So per your definition either you can’t influence the outcome in which case it’s gambling, or you can in which case it’s illegal and you’re committing a crime.

All that said I think the definition of gambling being truly random isn’t quite right. Gambling has always been about statistical outcomes because if it was truly random nobody would ever make any money off of it so nobody would waste their time. Remember the house always wins because they control the odds. These sites are just bookies in techbro clothing using marketing and legal loopholes to run illegal gambling sites.