bart_spoon

Seems like that only applies if you anthropomorphize the AI and consider it the author, rather than a tool utilized by the artist. I mean, yes, the AI is doing the bulk of the work, but so is photoshop for a lot of digital art.

Ok but who is actually saying the content/product result isn’t important? Seems like a strawman.

And yet isn’t Pixar famous for their intensive, structured process for movie development?

The opposite is also concerning. IP law has always been convoluted, messy, contradictory, and morally ambiguous. The complaints of IP violation by LLMs are simply taking these inherent flaws and making them immediately obvious, forcing decisions that ultimately will set precedents on the legality of human thought that I don’t think anyone will be comfortable with. People understandably see OpenAI and Microsoft as potentially dangerous to be given so much leeway, but fail to consider on the flip side companies like Disney who have already more or less dictated the majority of copyright law for decades now. They must be chomping at the bit at the legal precedents potentially coming down the pipeline that call into question the ability to interact with any kind of media or information at any level without potentially being on the hook monetarily.

I think all this is doing is making us realize that we have built a massive economic system on a fundamentally flawed idea of ownership over ideas, and the only two solutions will be to tear up the rule book, which will be extremely painful, or double down, which will be fatal.

The concern many MtG players have is that throwing a bunch of other IP crossovers into the card game has a variety of potentially negative effects:

- established players who enjoyed the theme of Magic in the first place might be turned off

- it creates an increasing number of products and therefore product fatigue, which is something I’ve heard lots of players mention and was even brought up by Hasbro themselves in a financial call

- it adds to the increasingly confusing legality of cards/formats. Some crossovers have alternate in-universe versions. Many don’t. Some are commander legal. Some are commander and modern legal. Some have their own draft environments. Knowing what cards you can play in a given format is getting obnoxious to track.

- They clearly seem to be introducing new players to Magic, but they are also clearly driving old Magic players away. The question is how persistent are the new players? Do they stick around, or does their interest wane when their favorite crossover franchise isn’t getting new cards anymore?

I personally think it’s too soon to tell, but a lot of the moves strike me as very reminiscent of the types of moves that eventually are identified as enshitificafion as management tries to extract as much short term juice as possible in a way that damages long term health. I’d love to be wrong.

There’s only two camps, investors and gamers. I think that anyone who wants to “have cards that have relative value” are investors, whether they realize it or not. Its one thing to collect because you like something inherent about the cards themselves (I recall a recent Reddit post where the user was trying to collect every single Magic card that depicted an owl in any way). But if your concern is about the monetary value of the cards in your collection, you are an investor, of some kind.

I personally think cards should be for playing, and am pretty opposed to cards being valuable if that means that playing the actual game becomes prohibitively expensive. Standard decks costing hundreds of dollars is not a good thing.

“Cure” is a tough bar, but I believe Paxlovid, the anti-viral used to reduce Covid severity, was identified using ML. There’s many companies like Recursion Pharma which are entirely focused on using ML for drug discovery, and from what I can tell seem to have promising results, but drug development is slow enough that nothing will come of it for a while.

Also, while not medicine focused, Google’s GNOME project results announced a few weeks ago was pretty remarkable. They discovered more theoretical new materials using their ML approach than the rest of human history combined, and they are already confirming many of the results in laboratory settings. That has the potential to be a revolution in limitless scientific and engineering applications.

It’s the cost of compute hardware required to train a model of that size

Understanding meaning what?

DBT is open source and can be used entirely for free. They sell a hosted version.

Also DBT and Airflow are different tools, aimed at solving completely different problems. The Airflow competitor in this article is Dagster, which is also open source.

Far more helpful to me than these essays on the benefits of a particular paradigm is a simple, minimal example in a repo that I can dig into and explore.

Reading through this entire exchange and it would seem your argument inevitably falls back on this argument of consciousness, which is both arbitrary (why does that matter in the original context of the distinction in learning, particularly from copyrighted works?) and ill-defined (what even is consciousness? How do we determine if another being is conscious or not?).

Feel free to expound on the important differences

> For example any discussion about copyrighted works, which is a hot topic, will inevitably end up with someone equating an LLM "learning" to a human learning, as if the two are identical.

I don’t think the point is that the two are exactly identical or that humans and LLMs are equivalent, but that the processes are similar enough in the general level that any attempt to regulate LLM training in copyrighted material will inevitably have the same ramifications for human learning.

In pretty much all attempts I’ve seen to differentiate the two, it inevitably boils down to hand-waving about how human beings are “special”.

That is an oft repeated myth. A recent study by UCSF of California homelessness found that 90% of homeless Californians became homeless while already living in California, and 75% still live in the same county as they did when they became homeless [0]. Locals are approximately 90% of the problem.

[0] https://amp.theguardian.com/us-news/2023/jun/20/california-a...

Altman’s departure represented a real possibility they may miss out of enormous amounts of wealth. People are willing to put up with or even support a lot of things when money is on the line.

I think a non-LLM model would be good for learning to draft within a particular set, because it can simply encode each card and learn what cards do well together, comparative strength, etc. without learning the stats/rules text on the card.

For learning to draft in one environment and then applying it to a new set of cards, like done here, it would become far more difficult to do without an LLM. The variations and nuances of cards rules text would have to be encoded as features, which would be extremely cumbersome. The LLM gets you some level of understanding of that for free.

This has been true for a lot of publications I used to like to read. Wired, Time, Sports Illustrated all are much, much worse today than I remember them being 20 years ago. I don’t know if that’s me or a genuine change in their quality.

One doesn’t need arms and legs or any physical presence at all to do immense damage to society. Ransomware attacks cripple companies, hospitals, utilities, and governments on a regular basis. The US and Israel have demonstrated, in at least one instance, the ability to create computer viruses that can cause industrial hardware to physically destroy itself.

Our world has been so completely integrated with digital technology that if it were disrupted in a large scale fashion it could easily result in millions of deaths. No power, no internet, and people lose access to heating/cooling, clean water, hospitals can’t run, food production is disrupted, money can’t be accessed, medicine can’t be produced, communication is cut, shipping and transportation of what goods do exist grinds to a halt.

And that’s just in the world of malware, that doesn’t consider all the ways it could just inject enough discord into society that it doesn’t need to physically attack us, we will do it ourselves.

So according to you, human beings are considered to possess intelligence prior to grade school?

Many children (and adults for that matter) would fail to qualify according to that definition of intelligence.

Transformers came about in 2017, so extend the time horizon by a whopping 2 years and OP’s point remains.

AlexNet was 2012, which is what essentially kicked off the deep learning neural network revolution. The original implementation of AlphaGo utilized DNN in 2016 to beat Lee Sedol at Go, which people had been predicting was at least 5-10 years away.

So the timeline, spanning just under the last 12 years at this point: AlexNet -> AlphaGo (4 years) -> Transformers (1 year) -> GPT4 (6 years)

Imagine showing GPT4 to someone in 2012. They’d think it’s science fiction. The rate of progress has been absurd.

The conflation of concern over AI safety and effective altruism is bizarre and unwarranted in my mind. Just because there is overlap between adherents of the two does not make them the same.

Effective altruism largely seems like billionaires attempting to justify their insane wealth. AI safety is people simply understanding their work doesn’t exist in a vacuum and recognizing that technology has had major effects, both positive and negative, on society.

Yes, maybe AI doomers will be wrong some times. But I’d prefer people bringing about world-changing technology to genuinely care about the potential negative consequences of their work and act accordingly. I’m surprised that there are individuals who don’t.

Isn’t that reference like the exact opposite? It’s nominally self-aggrandizing (he’s the only one who could have gotten it right) but is actually about a higher goal (sacrificing himself for everyone else’s sake). That’s how I remember reacting to it anyways.

Stock options?

Microsoft has already done major layoffs over the last year of their own employees. Why wouldn’t they lay off OpenAI employees?

That is entirely dependent on how that wealth is obtained

Perhaps because, for all of Silicon Valley and the tech industries platitudes about wanting to make the world a better place, 90% of them are solely interested in the fastest path to wealth.

Was it due to incompetence though? The way it has played out has made me feel it was always doomed. It is apparent that those concerned with AI safety were gravely concerned with the direction the company was taking, and were losing power rapidly. This move by the board may have simply done in one weekend what was going to happen anyways over the coming months/years anyways.

> The exchange of services and goods in a market is positive-sum.

As long as you ignore externalities, yes.