dougb5

This makes me think about the credibility of single-author vs. multi-author papers in different disciplines. In computer science, a paper is seen as suspicious if there's just one author (at least nowadays). But in economics it seems much more common. Can an economist explain this for me (or perhaps a paper written by multiple economists?)

My cynical answer: I think they reframe their history to see it as a series of ever-larger successes -- with some setbacks maybe. They avoid introspection that might lead to the unpleasant conclusion that they have actually gone off the rails and caused societal harm since their initial success. By and large their net worth has continued to increase, and that's the measure that they're comfortable with associating with success, and they see that as validation. Plus, they are surrounded by people who have the same incentives to lie to themselves and to each other.

Further, the people who have purchased this situation for us, and the wealthy tech VCs who are staying silent now, are all people this community once admired.

Even if the VC-backed companies jacked up their prices, the models that I can run on my own laptop for "free" now are magical compared to the state of the art from 2 years ago. Ubiquity may come from everyone running these on their own hardware.

What exactly should be rate-limited, though? See the discussion here -- https://news.ycombinator.com/item?id=43422413 -- the traffic at issue in that case (and in one that I'm dealing with myself) is from a large number of IPs making no more than a single request each.

This matches exactly with what I'm seeing on my own sites too and it's from all over the world, not just China.

(I described my bot woes a few weeks ago at https://news.ycombinator.com/item?id=43208623. The "just block bots!" replies were well-intentioned but naive -- I've still found no signal that works reliably well to distinguish bots from real traffic.)

Sutro Tower's height is boosted by the fact that it sits on Mt. Sutro. It reaches 1,811 feet above sea level, just higher than the tallest building in the US (One World Trade Center in NYC, at 1776 feet).

I like to mention this fun fact to my visiting friends from New York, conveniently overlooking any higher-elevation buildings in Denver.

I sure don't agree that constitutes extreme activism of any kind, but anyway, my question is for Ancapistani and it's about the personal friends referenced in their message.

> They are also some of the most vocal extreme left-wing activists I've known well.

Can you give a few examples of the vocal extreme left-wing activism?

When you say "big hitters", I guess you mean the well-known corporate crawlers like GPTBot (one of OpenAI's). Yes, these do tend to identify themselves --and they tend to respect robots.txt, too -- but they're a small part of the problem, from my perspective. Because there's also a long tail of anonymous twerps training models using botnets of various kinds, and these folks do not identify themselves, and in fact they try to look like ordinary users. Collectively these bots use way more resources than the name-brand crawlers.

(My site is basically a search engine, which complicates matters because there's effectively an infinite space of URLs. Just one of these rogue bots can scrape millions of pages from tens of thousands of IPs; and I think there are hundreds of the bots at any given moment...)

Diabolical! But I think the occasional humans that I falsely identify as a bot will not appreciate this.

That is helpful, thank you! I will check it out.

Oh, I've tried captchas, and I can say that they're an awful experience for the humans you accidentally hit, at least for a service like mine that is relatively low value per session (a dictionary website). Within minutes of changing my WAF configuration to captcha users that it thinks are high-probability bots, I'll get angry feedback from users in Singapore (say) who don't want to have to solve a puzzle in order to look up a word in the dictionary. I don't blame them.

I like the Cloudflare challenge ideas suggested on this thread, though, I might try them again.

I'm on AWS and use their WAF service to do some rudimentary bot blocking, but CDNs (their Cloudflare equivalent, Cloudfront) have been too expensive in the past and the bot control mechanisms they offer have too many false positives. Perhaps I should check it out again.

Part of the problem is the economics of it -- I've chosen to self-fund a high traffic site without ads, and that's on me. But it was possible to do this just a few years ago.

AI is killing my website but in a different way than what's discussed in the article: Content scrapers are completely out of control, spiking my serving costs and degrading performance for human users. There seem to be hundreds of them and they've gotten very good at looking like human users so they're hard to block or throttle. I can't prove they're all AI-related scrapers, but I've been running the site for 25 years and this issue only became problematic starting in, oh, late 2022 or so.

USA Today has had something like this for a long while [1], though whether they find the strongest voices on each side is debatable.

[1] https://www.usatoday.com/story/opinion/2018/01/26/why-does-u...

The a16z article starts with "View this report on Gamma", so I'd assume they're on board with it. Maybe they're an investor?

> Voice agents allow businesses to be available to their customers 24/7 to answer questions, schedule appointments, or complete purchases. Customer availability and business availability no longer have to match 1:1 (ever tried to call an East Coast bank after 3 p.m. PT?). With voice agents, every business can always be online.

I don't get it -- textual support chatbots have been around for decades. Even if we accept the premise that people would rather speak to them by voice, how do voice agents represent some kind of sea change in availability?

(And I personally find customer support chatbots deeply frustrating to use for reasons that have nothing to do with the modality or the quality of the AI model. I only ever need to use one when the question I have is not answered in the documentation, which is often the extent of the chatbot's business-specific training data. Inevitably I end up being led in circles, screaming for a human.)

Could you change your product to only show images that have Creative Commons licenses? https://daminion.net/articles/tips/how-to-use-images-legally...

My complaint is with the article. They haven't done the reporting needed to tell me how much US taxpayers actually spent on this. They are happy to leave readers with the impression that we funded all of it. This is how we get distortions like "USAID funds Politico!" spread by bad-faith actors.

The other questions they could have answered: How much did USAID spend on IFPRI? Was the grant from USAID to IFPRI expressly designated for v-Fluence, or was it for other programs as well?

> part of an effort that was financed, in part, by US taxpayer dollars

The "in part" here is doing a lot of work. I can't tell how big the part is from the detail provided in the article:

> v-Fluence, which also had the former agrochemical firm Monsanto as a client, secured some funding from the US government as part of a contract with a third party. Public spending records show the US Agency for International Development (USAid) contracted with a separate non-governmental organization that manages a government initiative to promote GM crops in African and Asian countries.

> That organization, the International Food Policy Research Institute (IFPRI), then paid v-Fluence a little more than $400,000 from roughly 2013 through 2019 for services that included counteracting critics of “modern agriculture approaches” in Africa and Asia.

If we count second-degree links like this, I bet we'd find that most organizations (including for-profit companies) are "funded in part by US tax dollars."

Does the naming scheme they've used for models so far suggest that they care about SEO?

I've never had a human personal assistant. I don't have a sufficiently "busy life", at least in the conventional sense. I appreciate that personal assistants can be useful for other people.

And I'm not knocking the idea of agents. I can certainly imagine other tasks ("research wedding planners", "organize my tax info", "find the best local doctor", "scrape all the bike accident info in all the towns in my county") where they could be a benefit.

It's the focus on these itty bitty consumer tasks I don't get. Even if I did have a personal assistant, I still can't imagine I'd ask them to make a reservation for me on OpenTable, or find tickets for me on Stubhub. I mean, these apps already kind of function like assistants, don't they?, even without any AI fanciness. All I do is tell them what I want and press a few buttons, and there's a precise interface for doing so that is tailored to the task in each case; the UX has been hyper-optimized over time by market forces to be fast and convenient to me so that they can take my money. Using them is hardly any slower than asking another person to do the task for me.

> The mouse and keyboard are definitely dying (very slowly) for everyday computing use.

Not to be that guy, but where's the evidence for this? People have been telling us that voice interaction is the future for many, many years, and we're in the future now and it's not. When I look around -- comparing today to ten years ago -- I see more people typing and tapping, not fewer, and voice interactions are still relatively rare. Is it all happening in private? Are there any public metrics for this?

I'm reminded of Kurt Vonnegut's famous story about buying postage stamps: https://www.insidehook.com/wellness/kurt-vonnegut-advice

"I stamp the envelope and mail it in a mailbox in front of the post office, and I go home. And I’ve had a hell of a good time. And I tell you, we are here on Earth to fart around, and don’t let anybody tell you any different...How beautiful it is to get up and go do something."

> We’re collaborating with companies like DoorDash, Instacart, OpenTable, Priceline, StubHub, Thumbtack, Uber, and others to ensure Operator addresses real-world needs while respecting established norms.

Are these tasks really complex enough for people that they are itching to relegate the remaining scrap of required labor to a machine? I always feel like I'm missing something when companies hold up restaurant reservations (etc.) as use-cases for agents. The marginal gain vs. just going to the site/app feels tiny. (Granted, it could be an important accessibility win for some users.)

Wow, thank you, that means a lot to me! And I'm stoked you found Spruce -- we added that 5 years ago, pre-LLM era, but it never got much traction.

Allow me to plug a dictionary/thesaurus site I've been running for decades called OneLook (https://onelook.com). Although it's not specifically aimed at songwriters, it does attract many of them as users. Over the past thirty years, I've added all sorts of brainstorming features for creative writers—like the ability to search for words by description, to match words with a given meter, and more recently to discover which colors a word might evoke or vice versa.

(I’m also the creator of RhymeZone so I'll plug that too! I no longer operate it, but I can pass along any feature requests you might have to its new owners.)

We had fewer distractions, though, even if we didn't have all the world's information so close at hand. I think I'd pick that struggle again over what today's students go through. But I'm glad to have lived in both eras.