Hacker News Reader

Another test is whether kids use it w/o being forced to. My kids are obsessed with both MidJourney and ChatGPT. They have found use cases w/o any profit motive and w/o any outside pressure.

No amount of Blockchain "Webinars" would have convinced my kids to use the blockchain.

If blockchain did fulfill the claims of it revolutionizing the world of finance, it's still unlikely your kids would have become obsessed and regular users of it. If room temperature and pressure superconducting or nuclear fusion became viable, your kids might also be unlikely to play around with them on their own. Kids apparently love Roblox and most adults think the metaverse is dead. What kids like seems pretty orthogonal to what is legit or snake oil.

Point taken, but how about these:

- My mom cant figure out Blockchain either

- My sister (doctor) and brother (3 degrees) cant figure out Blockchain either, because they are too afraid of scanning the wrong QR code and having their balance drained

My mom, sister, and brother all use ChatGPT. My brother uses it actively for work.

You can dig deeper.... what exactly are the real-world applications of an immutable transaction ledger? I can't think of very many. Let alone a bunch of kids.

Decentralized is critical to the recipe as well. 'Immutable' centralized ledgers are largely a non-starter because the scenarios you want to use block chain for are zero-trust type scenarios. With centralization, the promise of immutability holds exactly as long as it's convenient for the centralizing power.

In any case, defining public use or accessibility as the metric of relevance of something is quite illogical. You end up with countless absurdities like relativity being worthless and CocaCola being the most relevant invention ever.

Immutable decentralized public ledgers could have many usages, but the biggest problem is: who is authorized to add new entries? There are different answers to this question depending on use case, but it is really the crux of the matter.

I don't blame them since the stakes are much higher, but it probably doesn't help to say "they can't figure out blockchain" because that doesn't make any sense.

Which 'blockchain'? Blockchain is a singular noun, that's like saying "they can't figure out phone".

Sure. I'm not saying Blockchain fulfilled it's promise. I'm just not convinced that every major new thing will be immediately adopted by kids. It seems like it's an orthogonal thing.

> it without being forced to

I think this is an important detail. A lot of "forced into it" numbers will get presented as "happy users" in various unfair metrics comparisons.

> Kids apparently love Roblox and most adults think the metaverse is dead.

Can you define what a metaverse is in this case and how is it different from what we used to call an MMORPG?

Roblox involves no block chain and no VR, so what exactly is "metaverse-y" about it?

Anyone can create and monetize their content in a world where a user's character can access all of the content. MMORPGs generally seem to be built by a single developer who controls the whole experience. They're also generally a single game instead of thousands of mini games.

> If room temperature and pressure superconducting or nuclear fusion became viable, your kids might also be unlikely to play around with them on their own.

I disagree. Kids will turn anything not too abstract into toys. An improvised radioactive sample (actually some neon indicator bulbs in a box) + Geiger counter brought by my parents = a version of a hide-and-seek game from my own past.

Also, a model maglev train is something so obvious that I am sure such toys will appear immediately once the technology becomes viable.

Not a great test, teens in the last few years sometimes get into crypto.

Only for the investment purpose by thinking they will get rich quick. I have never seen anyone using crypto other than that. Author specifically said not for profit.

Trading it on a centralized exchange to speculate on its future value doesn’t involve using it at all, it’s just a MySQL database at Coinbase tracking your bids and balances.

I’d wager the biggest non-speculative use case is pig butchering scams and ransomware.

> I’d wager the biggest non-speculative use case is pig butchering scams and ransomware.

Probably, but drugs are also a big use-case (and I don't just mean retail, its used in wholesale for transferring sums of money internationally to buy precursors, for example)

Plenty of kids I know use the blockchain, for their fake ids I think it is the backing financial infrastructure (crypto knowledgeable friend purchases ids and redistributes), for drug purchases

Midjourney and ChatGPT are both free to try, if someone gave me free bitcoin I might have made a wallet.

They had this in the early days. Bitcoin Faucet

Sigh... I "parked" my free bitcoin in the wrong place. Didn't know yet about just holding on to the private key and let it sit on the chain.

I'm routinely using LLMs in place of search engines. Would you rather type in a cocktail and instantly get a recipe? or do you want to open google, scroll past 50 ads, click through a bunch of blogspam that gives you the history of cocktails, the prohibition, the chemical formula for ethanol and their life story, followed by a recipe that was itself generated by a substandard AI model anyway

edit: I'm prepared for the enshittification of GenAI, but given competitive edge, many people will pay top $ for faster/higher parameter inference, I'm hoping ad support is minimal on the super deluxe god tier AI subscription which I'll have to pile all of my spare cash into.

If you pay for search you get some of those benefits already

Sincere question: what paid search do you have in mind?

How conceptually could a paid search engine solve the problem of SEO and spam? (Curation seems economically infeasible?)

Is what you're actually proposing ad-free search results? (I'd be willing to pay Google approximately $0.01 per search for this. I suspect this is probably more than they get from anything but the most expensive ads (pharma ads that confused old people will click on).)

What I would really love is a browsing-assistance agent based on something like an LLM that can summarize and de-cruft search results for me and show me content extracted from web pages in a standard format without any Javascript. (I don't need an LLM to be the arbiter of truth, I just want it to be good at summarizing documents.)

(Because Google is so wedded to the ad ecosystem I'm not sure I trust them to be the steward of technology that is basically about removing ads from my life.)

Kagi https://kagi.com/ is the paid search I've tried. It's was decent, but not a noticeable enough improvement to win me over as a customer.

OTOH, I am a paying ChatGPT Plus user and there are some days where I feel like I get my monthly subscription price's worth in a single day.

In terms of summarization, this study Anyscale just published is interesting - llama2-70b gets within a hairsbreadth of gpt-4/human scores for summarization, so it's conceivable that sometime soon, you might be able to run a local LLM on your browser and get decent results (4-bit quants of llama2-70b currently take about 40GB of memory, you need at least 2 x 24GB or 1 x 48GB GPU to run them at reasonable speeds atm): https://www.anyscale.com/blog/llama-2-is-about-as-factually-...

Who do I pay for better search? Honest question, I'd love a recommendation.

https://kagi.com/ is the one I see recommended. The results went to focus hard on "no blogspam" which is nice, and supposedly you can completely ban whole sites from your results forever (e.g. goodbye GeeksForGeeks and your shit SEO hacking)

I like Phind

I have two concerns when people tell me they're using this stuff like a search engine:

1. How expensive are these to operate when compared to search engines? If search engines are about as good at information retrieval, but are much cheaper, what does that mean for the future of these services?

2. How good are these systems actually when compared to search engines? I get that for things like code, you can immediately verify results through testing, but GPT was trained on what looks good to humans thanks to HFRL, not necessarily for correctness.

I really wonder what's going to happen with this comparison in the near future. I think there's two possibilities:

1. Running inference on models and keeping the models relevant by introducing new data remains extremely expensive and companies like OpenAI operate at massive losses. Eventually this leads to their products getting worse.

2. Models get cheaper to run and efficient enough to run on mobile hardware without losing much fidelity. OpenAI still might be in trouble in this scenario if "open source" models like Llama2 get popular.

> or do you want to open google, scroll past 50 ads, click through a bunch of blogspam that gives you the history of cocktails, the prohibition, the chemical formula for ethanol and their life story

I think in this aspect the difference between GPT vs Google not in level of tech, but because GPT took text from copyright owners and distributes it without their consent and without showing their ads, which is being challenged legally in my understanding.

Doesn't matter. If governments blunder this and try to enforce copyrights in AI, then we are all going to be using Chinese AI in 5 years instead of Google because they don't give a damn about copyright.

CCP already created its own regulation rules for AI.

And what does it say about copyright? Genuinely curious.

internet articles say something about copyright protection, but more important part is that AI has two promote core values of socialism and avoid statement criticizing Chinese government.

Interesting that the CCP rules for AI are the same as the rules for any meatI under their control. At least they are consistent.

Or you can run your own open source AI, on your own hardware.

I prefer search engines because I can look through multiple sources. LLMs just give an answer without context or what alternatives might be. They make it difficult to judge to validity of the response.

perplexity.ai usually gives 4-6 sources that you can validate the response with.

As far as I can tell, it performs a search and then feeds the full text of the search results to an LLM and present a summary with citations.

Not all questions need a variety of answers. If its good enough for your needs go for it. Saves some time searching.

It depends on how important it is that I get a precisely correct answer. For a cocktail it's not a big deal if I get one that's slightly different, in other cases an incorrect answer might be catastrophic.

Amusingly, Kevin Roose and Casey Newton made a ChatGPT-authored cocktail recipe on "Hard Fork" a while back: https://www.nytimes.com/2023/06/30/podcasts/hard-fork-ai-poi....

I think the review was basically, "I’ve had worse cocktails. I’ll be honest."

So...meh?

Using bing chat, you always just check the source of its answer if it's crucial. Saves time searching.

The authors specifically state that they do not think that GenAI is snake oil. You can’t legitimately argue that it is.

On the other hand, I’m not sure that you can legitimately argue that it is not over-hyped. The only way that it isn’t is if we achieve “singularity” imminently - because that it is what a sizable number of people, including possibly some people at OpenAI, are actually expecting.

I think we’re nearing the point of the initial dotcom bust. A whole of internet startups went completely belly up when the early 2000s initial bubble burst.

That was no reflection of the actual potential it just indicated that people at all points of the value chain had not yet grasped how or what the internet was or how it would mesh with humanity.

That’s where we are now. It’ll burst but not because it’s overhyped; because we don’t collectively understand the implications.

Snake oil is about the thing being sold, it's about the sales pitch. It's the people selling the product.

An important snake oil trait is making grand claims which have little basis as fact, particularly with things which are difficult to understand or difficult to verify (if not impossible.) You can say anything about a product if the prospective buyers can't verify the claims. And so, the claims tend to get broader and flashier.

Therefore, a snake oil sales pitch could still be applied to a highly useful product.

Are people durably achieving a competitive advantage through the GPT?

I think we've seen some people striving for a competitive advantage, and willing to see one exist, but I don't know if we've seen any "AI company" turn a meaningful and durable profit yet (open to being wrong on this).

To me (so far), it seems like GPT use cases that apply it as a table stakes feature of some greater application (rather than an ecosystem or a platform), are the ones that are actually showing promise in terms of expected utility + value capture.

If GPT4+ level engines become as cheap to execute as a SQLite query, I think that's where things get interesting (you can start executing this stuff at the edge).

But I still can't see new companies (a la OpenAI, MidJourney, etc.) making a lot of money in this scenario, it seems to overwhelmingly favor companies that already have distribution.

For programming it’s like an additional feature on top of Google. You can dump entire error logs, get correct syntax (i give it a 4/10 for accuracy though), and even dump your entire page of code if there’s a bug in it.

Yea the sql queries, string or array manipulation are better and fast than google 8/10 times because of the errors it generates.

We're likely way too early to call this one. Based on current hardware trends, GPT-4 will run on your phone within 6-10 years. Right now, folks are feeling out the edges of what makes sense and what doesn't make sense at extreme R&D and opex expense. In 10 years, we'd expect the winners of today to still be winners and have great margins due to declining compute costs.

Granted, if you are spending 10x the future value of the product your are offering ... then even a 10x decline in compute costs won't get you where you need to be.

> Based on current hardware trends, GPT-4 will run on your phone within 6-10 years

they quantized model from 16 bits to 4 bits which was low hanging fruit, and looks like they can't quantize it anymore to 2 bits..

CPU/GPUs are general purpose, if enough workload demand exists specialized Transformer cores will be designed. Likewise, its not at all clear that current O(N^2) self-attention is the ideal setup for larger context lengths. All to say, I'd believe we have another 8-10x algorithmic improvement in inference costs over the next 10 years. In addition to whatever Moore's law brings.

Mobile TPUs/NPUs:

Pixel 6+ phones have TPUs (in addition to CPUs and an iGPU/dGPU).

Tensor Processing Unit > Products > Google Tensor https://en.wikipedia.org/wiki/Tensor_Processing_Unit

TensorFlow lite; tflite: https://www.tensorflow.org/lite

From https://github.com/hollance/neural-engine :

> The Apple Neural Engine (or ANE) is a type of NPU, which stands for Neural Processing Unit.

From https://github.com/basicmi/AI-Chip :

> A list of ICs and IPs for AI, Machine Learning and Deep Learning

I agree with your last paragraph.

In my opinion, Microsoft's proposed 365 Co-pilot pricetag of $30 per month per user will probably make a profit. High usage users being offset by low usage users when corporate 365 group memberships come into the mix. Many corporates will take a chance on it, willing to throw plenty of money at anything that is perceived as having a chance of improving hard to define competitive edge, creativity and productivity.

Given the high-end art and luxury goods market has had a large use for money laundering, crypto targeting that market makes complete sense. That it happens to be illegal? Well that just requires some more ju$tification.

> Heck, my mom discovered it herself on Bing's homepage and was telling me I should check it out.

Sounds like grade A Microsoft marketing speak. But wait, there's more!

The predominate usage of NFTs are non-art based. Insurance policies and market making use NFTs very heavily, especially after a BSL license from one market maker expired this year.

And for art NFTs, much of the interest is from the collectors space or collector sentiment as these are an improvement in supply transparency and provenance in comparison to other mediums of collections.

Well that's a bad criteria. A lot of people get hooked on "snake oil", and will even try to seek it out. Astrology, alternative medicines, crystals, vibes, etc., cults and other such religious beliefs, various social, economic, political and other beliefs that aren't supported by evidence. Lots of them aren't "zealots", they're pushed by fear, misconception, misinformation, desperation, FOMO, hype, etc.

Also, even if some people in some usages get benefit out of it, does not mean that fuels hype and snake oil to try to sell it to uninformed people to use in inappropriate places. Baking soda is good for making bread, lots of people use it for that with great success. It is snake oil for curing cancer, however.

NFTs as domain names is novel, interesting, and non-financial.

https://ens.donains

My belief is that blockchain-based solutions are typically inferior to centralized ones except when the goal is to evade regulation. (Crypto is good for dealing in contraband, dodging currency controls, and collecting ransomware; but it's always less efficient than centralized solutions with a trusted authority. And the only time you can't find a trusted authority is when you're trying to dodge regulation.)

In this instance, I don't see how this offers any advantage over the usual registrar-based DNS: at the end of the day I'm just using the ENS rootnode keyholders as my registrar, and I don't see how the blockchain is solving any problem that couldn't be more efficiently solved with a traditional database.

"Evading regulations" is a phrase that holds a different weight when you don't live in a free society. "Evading regulations" could be using a gay dating app in Iran. Or organizing to fight against a repressive government in Russia. Even in so-called civilized Western nations, there are increasing tendencies toward autocratic rhetoric and encroachments into personal liberties. See: abortion access in the United States.

Not too long ago the US government put pressure on all financial institutions to stop processing payments for Wikileaks without them ever having been convicted of committing any crimes. That's just a single example plucked out of the sky, regardless of your feelings about that organization.

The great advantage that we get from a decentralized, opt-in trust infrastructure is that we have an avenue around these unjust and extralegal encroachments.

Having said all that, even if you don't care about politics, there are countless examples of centralized entities changing ownership, or changing strategies, and their users pay the price. See: Twitter taking people's usernames.

You might be fine with dealing with living at the whims of billionaires but I want my digital life to be more durable than that.

1. The first part here doesn't argue against my claim that "blockchain is only good for evading regulation", it argues that evading regulation is sometimes a good thing. That's fine, let's set that aside.

2. The second part argues about centralized entities changing ownership. I don't think the example of Twitter usernames is a good use-case for blockchain: there's nothing here that isn't already solved by public key crypto alone (which is how blockchain solves the problem anyway).

Maybe Twitter here is just an example and you want to talk about corporate control over payment services? I don't think this is a compelling argument. There's no shortage of payment services and even if one of them is bad you can always find another. Yeah: Stripe or your bank go out of business tomorrow, but switching to Apple Pay or a different bank is not a huge problem.

Corporate control over payments is a topic worth discussing but I'm talking about corporate control over identity. Twitter usernames are identities. Email addresses are identities. Phone numbers are identities. IP addresses are identities. Domain names are identities. All of those things are corporately controlled and can destroy businesses and communities on a whim.

I agree with you that public key crypto solves many of these problems, but you still need to publish your public key somewhere. And you need an infrastructure where public keys are first class citizens with the platform.

I gave ENS as an example precisely because it is open source digital public infrastructure where such things can be published forever, and it is not corporately controlled. It happens to use a blockchain to do consensus and create incentives for people to opt into running the public infrastructure. I personally think these game theoretical incentives are integral for the functioning of the platform but am not married to the idea if there are better ones.

Something like Ethereum is an anti-authoritarian platform. You don't use it unless you desire the qualities of anti-authoritarianism. For everything else there's centralized solutions.