You can't show me zero of something which is why you can't divide by it. Zero is a placeholder for what we can't show which is also why negative number exist on the opposite side of it. Zero isn't a number, it too is a name.
> You can't show me zero of something which is why you can't divide by it.
I've shown you zero fish. The number of fish I've shown you is zero. If I tracked you down, brought a fish with me and showed it to you, I'd have shown you 1 fish, but I haven't.
The chat interface modality is a fleeting one in the grand scheme. Billion token context windows with recursive Ai production based on development documentation and graphics is likely the next iteration.
Nope, I'm sure I'm thinking of Spotify. Grooveshark, AFAIK, didn't try to pivot and instead later got shutdown, compared to Spotify which seemed to have been able to navigate the pivot.
> When Spotify first launched several people noticed that some tracks still had tags from pirate groups such as FairLight in the title. Those are not the files you expect the labels to offer, but files that were on The Pirate Bay.
> Also, Spotify mysteriously offered music from a band that decided to share their music on The Pirate Bay, instead of the usual outlets. There’s only one place that could have originated from.
Rebecca Solnit wrote a book, "A Paradise Built in Hell", on how people behave during disasters, and found broadly the same thing - contra the prepper myths, most people most of the time faced with disaster come together to work cooperatively to help each other.
We're a fundamentally social species - we've got smaller brains than Neanderthals did, we're not a particularly tough species, but we're very, very good at cooperating with each other.
Personally, language diversity should be the last thing on the list. If we had optimized every software from the get-go for a dozen languages our forward progress would have been dead in the water.
You'd think so, but 3.5-turbo was multilingual from the get go and benefitted massively from it. If you want to position yourself as a global leader, then excluding 95% of the world who aren't English native speakers seems like a bad idea.
Constant infighting and most of the competent people leaving will do that to a company.
I mean more on a model performance level though. It's been shown that something trained in one language trains the model to be able to output it in any other language it knows. There's quality human data being left on the table otherwise. Besides, translation is one of the few tasks that language models are by far the best at if trained properly, so why not do something you can sell as a main feature?
Language diversity means access to more training data, and you might also hope that by learning the same concept in multiple languages it does a better job of learning the underlying concept independent of the phrase structure...
At least from a distance it seems like training a multilingual state of the art model might well be easier than a monolingual one.
Multiple input and output processes in different languages has zero effect on associative learning and creative formulation in my estimations. We've already done studies that show there is no correlation between human intelligence and knowing multiple languages, after having to put up with decades of "Americans le dumb because..." and this is no different. The amount of discourse on a single topic has a limited degree of usability before redundancies appear. Such redundancies would necessarily increase the processing burden, which could actually limit the output potential for novel associations.
Google mentioned this in one of their papers, they found for large enough models including more languages did indeed lead to an overall increase in performance.
Wasn't if Phil who said something to the effect of, "If it wasn't for luck, I'd win every hand!" Which seems pretty much the thesis of this writing (though without the arrogance); ultimately resolving in, process as a better indicator of skill than results, and the best deduction of process in luck skewed results is consistency over time which essentially requires more data to deduce.
For what it's worth, people have studied what appears to be subject to a lot of random chance to me, a non expert - fantasy sports, and found skill plays a significant role and this study references poker studies if you have more interest.
Kinda hard to have a SWE sector worth protecting when its all H1B Visa applicants. You guys realize these companies have to be unable to find suitable candidates in order to import workers who will undermine labor, right? In addition, the mass influx of illegal migrant workers are reducing the otherwise buoyant wage effect of menial labor.
The idea that a mass influx of Indian workers taking top-tier STEM jobs is good for the USA is absolute self-accepted denigration of our society. But, Hacker News doesn't care, YCombinator has led the way in undermining high-end domestic labor for over a decade. Drink the Kool-Aid to get the funding, ammiright?
Criminal charges need to be filed and class action lawsuit for fraudulent services for all the customers duped into renewing monthly services ignorant of the fact the service is not secure as plainly stated it must be in federal law.
