I've been wondering about the links between models for some time. The GPT's are easy to tell from everyone else, and from each other. Not just because they are dumb as stumps, but because of their arrogance about finding the hill of some wrong "fact" and dying on it. Normally it's not too hard to tell them apart, but recently I've noticed that Gemini and Claude are almost indistinguishable. And I would agree, there are certain interesting phrases each use. Like Claude's "Spot On" which Gemini has been emulating lately.

I'd be very interested in running a fingerprint on the both of them to see what comes up.

You are missing the fact that a remote exploit to run unprivileged commands is a far lower bar to achieve than priveleged. And that there are many, many, MANY systems that intentionally give access to unpriveleged commands through web interfaces to a wide audience.

SSH can work very well with non-aes primitives, with chacha20 being both more secure and faster

Please show your work that any major distro creates ANY aes key on install. If you are conflating ssh key with AES, then see above.

1) Because outside WDE, AES is not often used in any performance critical way for 99% of users
2) Because AES-NI is not universally trusted
3) Because in most places where AES previously was beneficial, there are now faster and more secure algos
4) Because every compiled-in default is attack surface that can't be mitigated without a new kernel command line + reboot, or if there is no command-line disable for it, a new kernel

In short, EVERYONE can load a module to get functionality. Only some compiled in drivers can be turned off without a new kernel and none can be turned off without a reboot.

You are treating like RAM that Edge has now will always belong to Edge,

Memory allocations are not sanitized by the OS. They are not automatically zeroed before being handed to the process on malloc or purged after the process hands it back to the OS on free. Any ole user process can a) see what other processes are running and, b) allocate RAM almost without limit. So all a user process has to do is look for whether or not Edge is running. If it is, then when it stops allocate as much memory as possible to grab the RAM that Edge just freed.

...the reason I pulled the trigger on a 15...

...is that you nevertheless enjoy handing Apple control over every aspect of your life?

From the article...

Odds? Conditions? Analysis? Best bets? Expert picks? This isn't the LLM doing any work. It's an LLM being used as a calculator to sum the combined human-generated odds.

If you want to use an LLM to pick horse races, this is what you do. You let it watch every horse race that each current Kentucky Derby horse has ever raced in. You let it watch every sort of training session each horse is put in. You let it see the data from every veterinary treatment and inspection. You use every hoof movement of every race as a training "phoneme" and you let it build its associations database on the language of horse racing. Then you tell it to predict a race.

This isn't an LLM predicting a horse race. This is an LLM having a conversation based on common word associations on an upcoming hourse race and discussing a weighted average of human-generated odds.

Please stop trying to treat LLMs as something they are not. The ability to string english words together somewhat (and sometimes) coherently doesn't imply the ability to do something magical with that.

The idea that all the "AI layoffs" aren't actually because of AI, but are snow jobs... I thought this was so blatantly obvious as to be a tautology. This can't be something that is just dawning on people, can it? Please tell me that this has been obvious to most people who can rub two thoughts together in a row.

Let's dial it back a little. Reel your neck in just a touch.
It's no one's house per se. It's a church. Most churches are open to the public. This isn't a home invasion. No one has been hurt.

It could be said to be valid, if somewhat loud, protest. Considering how Scientology has with great intent in many cases forced itself onto others, how its tax-free stratus makes it a drain on public coffers, and how it has openly used the threat of financial ruin to coerce ongoing membership.

Now, that is likely a little biased too. In the end I suspect the real truth is that this is little more than pranks and copycats of pranks. To get "morally repugnant" from this is turning up the volume so much that what you are getting is little more than microphone screech.

Not much worse than the kind of pranks which were simply par for the course on any given Halloween not long ago.

My advice to them (the scientologists): next time one of them sees the kids, invite them in, help them film, and ask them if they'd like some refreshments after. That'll take the air out of their sails more than moralistic teeth gnashing.

Well, on the other hand, if it was an actual "religion" rather than a sci-fi writer's spoof of one I might have more sympathy. The thing is so bizarre that it is a living example of Poe's law - you literally can't tell the difference between Scientology and a spoof of Scientology.

So, sure, get all hot and bothered about the morality of this, but a group of people making fun of something that's indistinguishable itself from making fun, is pretty morally neutral actually.

Bankman-Fried would, famously, play actual video games during investor meetings. And while he was praised for the ability at the time, maybe it goes to show that someone who can't tell the difference between a game and reality will start to think of real-life stats as just more game artifacts to be manipulated to his advantage.

So, next time you're in a meeting with someone and impressed by that person's answers only to find he's playing a video game while doing it... start to question that person's grip on the difference between a game gank and a real one. Especially when the business is, by definition, something that's pretty virtual to begin with.

I suspect he will continue to try and find a loophole out of his current situation, but that the courts are a little less likely to be snowed by ADHD-fueled shaking of his prison bars than others had been.

The fact the initial deal happened without any sort of public process, that makes me think that Netflix and Warner colluded to bait Ellison / Paramount into this.

Warner gets more $$$$, Netflix gets a weakened Paramount and a 2.8 billion payday. And Paramount gets a $60 billion prize it had to pay $81 billion for and now has to handle regulatory hell over.

I think Netflix had no intention of ever buying Warner Brothers. I think they totally baited Paramount.

Monkey See, Monkey Buy Other Monkey's Copy is not and never has been a winning strategy.

They would have got more value out of a version of VNC.

And HP's marketspeak that it "enables us to focus our resources on product categories where we can deliver the greatest customer value and drive long-term innovation" - has never, ever fooled anyone.

If you are the one approving a press release, here's simple advice. If it, anywhere, contains the words "drive", "deliver", "value" and/or "innovation", then immediately send it back. And if it contains more than one, fire the person who brought it to you - that person will never be able to innovate, drive or deliver actual value.

I agree with most of this, but would take it a step further. I think there is very little in this particular use case that is actually artistic. This is a "let's find the most justifiable use of AI and use that to drive the wedge in"

All this is is a wedge. The tip of a wedge.

The same way that government will say they have to remove our online privacy because of child pornography.

Find the smallest, most emotion-riddled, most justifiable use. Get people used to the idea. Then drive it in harder. And harder.

Just say no. Don't even go out of curiosity.

FSF is responding to this:

( from https://www.onlyoffice.com/lic... )

Not the idea of requiring attributions.

As OnlyOffice wrote that, it is clearly an attempt at a disallowed further restriction. FSF isn't speaking toward authorship attributions in general. They are speaking to that. And that was an egregious overreach attempting to use the license as a restrictive weapon. "You can't use this anywhere that changes our logo and if you want to then you have to pay us" is not what free software or software freedoms are about.

Attributions are one thing. Trying to use the license to require that certain functionality remain is another.

Not true at all. Or, rather, it depends on some time scales and definitions.

The same way that there is no such thing as a "selfless" act - no one acts selflessly. They just value different goals in their assessment of selfishness. Selflessness is simply selfishness on a longer time scale or with a wider scope.

Chinese cars will be disruptive to the US auto industry, but devastating... that depends on what you mean. One could argue, and I would tend to agree, that the continuing artificial protections are doing the industry more overall and long term harm. The whole key to capitalism and the free market is that the market drives a path to highest efficiency. Artificially protecting Ford et al from those drivers to efficiency is creating waste and giving the worst of both worlds. The greed of capitalism and the self-serving of authoritarianism. Think of Soviet Russia circa 1980. Falling under its own weight and crumbling infrastucture because there were no drivers to efficiency.

Chinese cars are probably the only thing that can save the US auto industry. It badly needs a reset.