peterlk

There's a lot of negativity in the comments here, and many of them have merit. However, the thing that is interesting to me about OpenAI, AI21, Cohere, and all the other LLM providers is that they are broadly useful, and often helpful. Perhaps they don't live up to the marketing hype, but they are still interesting.

For example, I used to have a biology blog, and I've been thinking of starting it back up again. I've been using OpenAI and Mantium (full disclosure, I work at Mantium) to generate the bones of a blog post so that I have something to start with. Coming up with ideas for my biology blog posts was almost 50% of the work.

If you're interested in judging the quality for yourself, I have a biology blog post generator here: https://f0c1c1e0-f6b6-46bc-81a1-eff096222913-i.share.mantium...

and a music blog post generator here: https://8aaf220e-4aff-4d4e-ae61-90f08011c9ac-i.share.mantium...

(they were both "created today" because I moved them from our staging environment)

I've been having "AI" debates like this for about 10 years now, and I think they usually go in 1 of 2 directions:

1. We don't know what intelligence is 2. AI can never be intelligent because humans are special (in various ways)

Of the two, I think that 1 is the more compelling to talk about. Let's look at state of the art Large Language Models (GPT, BERT, BART, T5, etc.) Everyone claims that they can't be intelligent because they're just cleverly predicting the next tokens. The most common failure mode of this is that they hallucinate - if you ask them to do something for you, they'll get it wrong in a way that kind of makes sense. There are some other more subtle problems as well like common sense reasoning, negation, and factuality. We could say that because of these problems they are not "intelligent". But why is that so important? Can we say with certainty that human intelligence is more than just patterned IO? If it is just highly tuned patterned IO with the environment, perhaps we have discovered intelligent systems, but they're handicapped because they're limited in their sensory perception (read: data modalities). And perhaps by combining several of these models in clever ways, we will end up with an architecture for pattern IO that is indistinguishable from human intelligence.

The naysayers claim that this won't work because we'll still end up with mere pattern prediction machines. But this starts to look like a "humans are special" argument.

I've just spent a bit over a year in the live entertainment industry, and feel like I might be able to provide some insight here. On the one hand, this feels like an own-goal because no one in live entertainment has been paid since the beginning of 2020. Though that started to change a little over this summer.

On the other hand, many of the people from the live entertainment industry found refuge in film and television. There has been a hollowing out of the industry where the senior folks have stuck around, but lots and lots of people who were early in their career have left and are not coming back. The senior folks can afford a few more weeks of no pay. However, if they cut off revenue to the film&tv world, they are setting up a 1-2 punch to the entertainment companies (1. COVID, 2. strike), and that puts them in a good negotiating position.

Anecdotally, many emigrants (a dozen or so) that I know are now choosing to go to Canada. The US has lost a significant amount of its shine over the last few years for them, and their skilled labor is going elsewhere.

I think there's a pretty critical piece of analysis missing here: inflation. An alternative explanation might be that housing prices are being driven up by inflation, and wages haven't caught up yet. There are two ways to lower the home value/income ratio:

1. Decrease housing prices 2. Increase wages

We haven't seen meaningful wage increases for a long time now, so perhaps it is time. Maybe this is just a symptom of wages not tracking with inflation, and the solution is to pay people more.

Sounds like you should read Hyperion! Such a good series...

How can I implement my website such that it is optimally useful for browsers like this? Just avoid flashy stuff? Is there any way to list my blog somewhere as a friendly place to read? I'm spinning it up again after a year off for reasons that I will probably write about :)

For anyone else who looks at the comments first, there is nothing to see here. I was hoping for a link to the original report, but this article just states that some VP didn't want an internally generated report to go out because the top shared content wasn't a good look.

According to this article, I think I rank as an optimist - I tend to trust people, feel that we can solve the problems before us, etc.

But I'm listening to a book right now called The Ascent of Humanity, and it makes some interesting points about the philosophical axioms of the arguments presented in this article. Namely, the argument is that the technological approach - inventing our way out of problems - is doomed to failure over the long term because it creates an accelerating treadmill of required progress in order to solve the problems it creates. This is reflected in economics by the requirement that our output will continue to increase (otherwise debt doesn't work).

I now look at articles like this a bit more skeptically because they assume that technological progress is necessarily good. For example, what argument could be made that painted longevity into one's 90s as a bad thing? Well, we spend an enormous amount of human effort keeping people alive for the last few months of life. At what point is it not worth it to keep people alive any more? Perhaps culturally, we need to become better at coping with and accepting death.

I haven't made up my mind on this, because the case against technology leads to some pretty bizarre conclusions. But I find it worthwhile to consider.

Since no one else has mentioned it, I feel compelled to mention thermonuclear catsplosions.

My understanding of the legend is that a player began experiencing a catsplosion, and wanted to find a way to get rid of the cats. So he started tinkering with the game a bit, and eventually tried setting the blood temperature to a very high number. This killed the cats, but also had the unfortunate side effect of setting them on fire. And the cats were multiplying faster than they were dying, so there was a massive, expanding fireball of burning cats.

If this legend is wrong, I'd love to hear another version.

http://dwarffortresswiki.org/index.php/DF2014:Catsplosion

We are currently in the "information superhighway" era of the metaverse. Everyone wants to have a proprietary metaverse that looks pretty much like IoT visualization or an open world mmo. What we need is a "browser" and common protocols. The biggest impediment to the metaverse at the moment is the size of 3D assets. A "website" in the metaverse looks very much like a game, but games are really big. There are solutions to these problems, but there's a lot of engineering work to do between here and there.

A taxi driver in Singapore told me that the highway in Singapore is also designed for planes to land on. My own observations indicated that this was plausible.

Epic Games has a $1B investment for the metaverse

Nvidia has Omniverse

Roblox hit a $58B market cap in June

Unity has VR chat

And there are a zillion startups/indie games that claim to be "The Metaverse". Of course, for the metaverse to be successful, we need a common protocol, not walled gardens - but this is deep topic, and I don't want to type it all.

Media is moving to the metaverse - Fortnite concerts, Roblox concerts, Porter Robinson's virtual event, Tomorrowland, Lost Lands (couchlands)

Enterprise is moving to the metaverse - Microsoft/Epic games partnership for global 3D info (can't remember the name), Nvidia's Omniverse, just go google search "digital twins"

Rocket League, Fortnite, and Roblox are partnering with big brands

The metaverse is taking shape before our eyes. What we're missing is a common interop format, and the leading contenders are USD (Pixar, Nvidia) and gltf (Epic Games, Mozilla).

I'm building a business for the metaverse at the moment, and there is something real if you look past the hype. The problem is that it doesn't quite look like the metaverse of _Ready Player One_ or _Snowcrash_. So it's easy for people to dream in the wrong direction. As you might expect, much of this is pretty incremental progress, but that doesn't make as great of a narrative, and the conclusion of the available incremental progress is monumental - but there's a ton of work to be done.

I love talking about this stuff. If you want to work with us, or just trade thoughts, feel free to reach out to me about it.

I similarly think that you are applying a techno-utopian lens to your analysis. My family and I have spent time with many people who did not have internet or other modern amenities (Mongolia, Mali, Burkina Faso, Kazakhstan, and several others), and your statement does not hold. It is not _entirely_ untrue, because a bacterial infection can kill you, however, the dynamics of life are just... different. People are still capable of all the emotion and fulfillment that you and I are, but life is slower, simpler, and more well-defined, so it is often easier to feel happy and fulfilled.

This brings to mind a conversation that I've had several times over the years. It goes something like this:

> Wouldn't it be cool if we could have code that would test itself?

> Yeah, if only there were some kind of automated way to encapsulate the requirements of classes and methods that could ensure that they behaved within some known boundaries...?

We've had automated unit tests for quite some time, but they are more often called compiler checks. Want your "automated unit tests" to be more comprehensive? Improve your type system and your compiler!

With all that said, I do not want to be dismissive of peoples' projects. This is fun and neat.

Meta: I'm really happy that there is a non-english article on the front page of HN. Can we make this happen wore often?

"Hire slow and fire fast". I've never seen a company that had not made several bad hires. The difference was the speed at which they were willing and able to identify them and remove them. You shouldn't trust someone the moment you bring them on, but I do believe that should be able to imagine yourself trusting them.

A general rule for me is that you should know within the first 2-4 weeks whether you've made a bad hire. This is a bold statement, but my experience is that this time frame will give you enough time to get a gut check (super vague, but there are too many different patterns for me to articulate here). If your gut says no, that will stick with you, and your gut will likely become a self-fulfilling prophecy anyways.

I see this argument a lot, and I think it is more hollow than it appears. I am certainly not the same person now that I was 10 years ago. Indeed, I am different than the person I was 1 year ago. Life happens, people change. Directing this change with psilocybin seems like a reasonable choice. Sometimes it may misfire, and you will be a bit worse off; but sometimes humans get depressed, angry, or otherwise negative for lots of other reasons.

I watched a movie once that made me deeply depressed for weeks. Thin may have permanently changed my personality too. Should we ban movies because they make people feel and think things that they haven't felt and thought before? (That's as close as I'll get to cancel culture uproar in this post)

> Nuance has largely failed to fully embrace deep learning approaches.

This may be a feature, not a bug. Explainability is very difficult with deep learning systems, and that often makes deep learning less suitable for some compliance applications like government usage. Microsoft may also still be feeling more conservative after their twitter bot learned so well a few years ago.

> you don’t need fairy tails to be a good person and to come to terms with death

I don't think this is true in general. I have a pet theory that much of the conspiracy theory following we are seeing is actually a weakening of church. People need spirituality - it doesn't matter where that spirituality comes from (except for local social pressures - but those aren't relevant to this point). If they're not getting spirituality from church, it has to come from somewhere. So people latch on to conspiracy theories because it provides higher purpose in their lives.

If you are teaching your children to be good and to not fear death, you need some philosophical backbone with which to unify these ideas. You may well be able to provide that, but not everyone is. Religion serves as a common philosophical substrate for discussions that people don't know how to have. It is imperfect, and definitely not as good as thoughtful parenting in my opinion, but it fills some of the gaps.

When I was in college, this video really stuck with me: https://youtu.be/FxF-mxuv4uI

The relevant quote (and my favorite quote from it) is:

> Stop trying to be Mark Zuckerberg, because the best you can possibly do is second place.

Huh... This actually seems like not a bad way to ensure that Google is aware of the problem. A sufficiently large reddit community could probably knock a lot of big fish off of youtube with this approach

Absolutely. I cleaned (removed cursing) more hip hop tracks than I care to remember with the highlight/reverse in Audacity. There is no other program that allowed me to do this so simply and efficiently.

One of my favorite quotes because it is so hard for me to swallow

> People will choose convenience over quality any day

- CEO of iHeart Media

I am neither a fan of iHeart Media nor that idea, but I think it may be correct. If true, an explanation for Reddit's poor quality is that they made media consumption convenient, and user hostile behavior goes unpunished because there is nowhere else as convenient as reddit.

Kind of. I mean, sure, I think it would be really neat to be able to replace the L1 cache myself, but this would probably necessitate a less efficient CPU. Is it fair to fault Apple for using an Intel CPU which itself has no mechanism for internal repair/replaceability? I think it is if we were able to apply the same standard to every product (this is where the VAT idea came from); then the repairability becomes the aggregate of all the parts used, and it's the manufacturer's job to pick repairable parts, but how far do you go?

Does it make sense to lower the repairability score because I can't replace the L1 cache? If the answer is "no", then what is the difference between an L1 cache replacement and a CPU replacement on a fully integrated motherboard (like Apple's new computers)? The answer in my mind is one of convention. Perhaps the solution to this is to enshrine the Von Neumann architecture as a legal standard. But this solution seems like it could limit innovation.

I think this pretty well sums up the loophole. You have the right to repair... what? If I buy a macbook, do I get to replace the case? keyboard? These seem pretty reasonable. But what about the right to replace/repair the L1 cache? This is obviously a different story - it would be nice, but the part probably isn't manufactured by Apple (maybe it is on the new Silicon?). My point is that at some point, things start to get blurry, and it is in the manufacturers best interest to continue blurring these lines. I don't think we can reasonably expect the government to do a deep dive into every product category to work all of this out - what would this look like for cars, where safety is a concern?

I think having the law on the books is a move in the right direction, but I don't expect it to have teeth until te can figure out how to actually measure repairability. Maybe the EU could experiment with a supply chain approach like VAT, but for a repairability score? This also sounds fraught...

If there is anyone on earth who could reinvent the internet, it is Tim Sweeney. The metaverse is coming, and it will exist because of him. Think I'm being hyperbolic? Go listen to his SigGraph 2019 talk

Plot twist: this is actually how Google decides to kill products.

Ok, well maybe this thread can be a tiny part of the solution to that stigma. First off herpes comes in two primary variants: HSV-1 and HSV-2. Most humans on Earth (between 60% and 95%) have HSV-1, and it's estimated that around 16% of humans have HSV-2. Herpes is extremely common. If you want references or more info, see the wikipedia page [0].

According to the article, the researchers are currently focused on HSV-1, and may work their way to HSV-2.

[0] https://en.m.wikipedia.org/wiki/Herpes_simplex

> Academic research shows that US multinationals consistently lead in adapting the best corporate management practices of any major country.[1]

Your source points to an article about horseshoe crabs... Can you update it? I'd be interested to red about your claim.