There are emergent effects of technologies getting cheaper and more accessible.
Surveillance was always possible but it was expensive because a person or two had to literally watch a suspect. So it was rare in practice and suspects had to be chosen carefully.
Mass surveillance is not new in the sense that surveillance was always possible. What is new is the scope of power it gives to those who can use it.
Not doubting the method works in general, but Simon Willison is a public-enough figure so the baseline level of info is higher than just HN comments. If you turn off Claude’s web search:
> Simon Willison is a British software developer, blogger, and open-source advocate, best known for…
Correct. And they usually don’t install windows on their computer either. The 8 year old laptop they got at Best Buy had it pre installed. So if Linux is going to go mainstream it’ll be because stores start offering PCs with Linux that are at a $140 discount.
> with blogs and email, we already had a decentralized social network
I’ve been using this social network lately. As someone whose inbox is mainly a dumping ground for receipts and confirmation emails, it feels kind of transgressive to write to and receive emails from a human being.
I have a blog and I thought about adding a comment section for substantive discussion. But that’s way more complexity, and lowers the bar for useless comments. Instead, I let other platforms handle it. And if people want to email me directly, they can and have.
I think we need to separate theory from practice. In theory, it can edit the training loop and come up with novel techniques. That is interesting.
In practice, the vast majority of the changes that auto research actually made would have been found much faster with BO if properly parameterized. You do not need an LLM to find a better batch size or learning rate.
I agree that many of the improvements found by auto-research systems could probably be discovered more efficiently by properly parameterized Bayesian optimization. Still, I think LLM-based heuristic guesses can be useful in some cases, especially for proposing reasonable initial hyperparameters based on prior knowledge reported on the web and in blog posts.
FPGAs won't rebuild fast enough for it to matter vs software simulation I'd wager. Even FPGA-in-CPU has been a dream for decades and there you have more time for some workloads, still never was commercially viable for general computing.
There was research a few years back that tried doing something like this with an FPGA, and they found that their algorithm actually exploited defects in the particular chip (not the model, the actual single specific chip) they were using to use electrical interference for computation that shouldn't have worked on paper. They could not reproduce their design on another FPGA of the same model from the same lot.
A year or two ago, would have agreed with you. I used to think they were cherry picking as well.
But Waymos have driven so many miles by this point, if they are hiding some data that would tip the scales back towards human drivers I have yet to see it. If there is a way to slice the data that makes Waymo’s look less safe I would welcome the correction.
If Waymo truly has 80-90% fewer crashes in the conditions they drove in, then it still has policy implications for places like Phoenix that do have good conditions.
One, even if all police in the U.S. did start as slave patrols it is a textbook case of a genetic fallacy.
Two, your article discusses several origins of police forces in the US. In Boston it had nothing to do with slaves because Massachusetts was not a slave state when they created a police system in the 1830s. And since Afroman was raided in Ohio, also never a slave state, it does not make sense to carry over southern slave-catching history into modern police culture.
> The first publicly funded, organized police force with officers on duty full-time was created in Boston in 1838.
This is from your Time.com article.
Second, fugitive slave extradition was controversial in northern states and from your Wikipedia article several northern states even passed legislation to protect fugitive slaves.
And why would northern states spend their own tax dollars to fund police forces to capture slaves? It doesn’t make sense. They created police for public safety reasons in cities.
And even if none of that were true it still does not address the genetic fallacy. Just because some police forces started as slave patrols does not imply that all police today are inherently white supremacist.
A boss at my former job would constantly generate AI images and send them to the graphics guy for him to make 'adjustments' and no matter how many times the graphics guy explained that these images are much more trouble to clean up than for him to make a new one from scratch by hand, he seemed unwilling to understand and kept spamming slop at him.
I really felt for the guy the first time I was in a meeting and somebody had generated their own project roadmap recommendations. This type of behaviour introduces so much noise and time waste in the system, I would love to know how it shapes up next to the benefits.
Don't even get me started on people AI generating personal farewell notes for retiring coworkers or whatnot.
Stan Sedberry (the author) has had an extremely productive day today. He wrote this article and nine other AI related articles. All today. I am very impressed.
But in reality there’s only a handful of things people care about for your restaurant: what, when, and where. Put up your menu, put up your hours, and put up your location. And a phone number.
I specifically tracked this problem and built https://lleu.site to try and get businesses in my city off of social media.
Built a menu editor. Has a built in blog and image galleries. Events calendar and event posts. Has a single page simple mode and multi page editor. Contact form with message intake and forwarding. Easy UI that I don’t change underfoot every quarter so its consistent. Works on mobile and low powered devices as well.
Kept the monthly price low and I’ve done cold emails, mailers, newspaper ads, online ads.
Still barely any takers. Probably a bit of a branding thing. Maybe its something else.
Restaurants generally have low margins. What brand recognition do social media platforms have with an average person? How much does it cost a restaurant to list themselves on social media? That's your competition, widely-known and free.
You'd probably have better luck if you gave them an actual dotcom address, and you managed all of their online presence for them (Google Maps, Facebook, Instagram, etc.). But even then it's hard to compete with free, especially when your prospective customers are stingy.
Edit: I suppose you're also competing with Uber Eats. You can already see menus there, whether you want delivery or not.
"lleu.site" might not be the clearest in regards to what the service offers. It reads too nerd. Something like "easyweb.site" or "yourown.site" might better describe it.
Thanks. I am definitely injecting too much nerd. I just couldn't help myself. I do have alternate urls available but point taken. I should probably redo the branding.
My first reaction was that it was visually intimidating for a non-computer person. I went through the workflow of the demo and it was pretty easy, but I suspect most people wouldn't make it that far.
Appreciate the feedback. Would you mind highlighting anything in particular? I can take it. My main efforts went into the platform.
I admittedly stumbled through the landing pages. Wasn’t sure how far to take it. Most of my designs end up rather simple but I was concerned it wouldn’t attract people. Its terrible if its having the opposite effect lol
IMO the four designs that I saw as examples are not attractive enough. Especially coming from the editor's builder, they should make a stronger showcase.
Potential customers want to see the menu (or product range or similar), location, a couple of pictures. It's supposed to give useful and necessary information. The intended purpose is:
Before: your cafe does not have a website.
After: your cafe has a simple website where people can see the menu, hours, and location.
This tool accomplishes that, and looks fine.
It's not supposed to give the viewer an aesthetic experience so novel and surprising, subverting the entire paradigm of cafe menus, to leave the viewer questioning reality and rethinking their entire approach to life.
If your product works as described, I think it would be great for a lot of small businesses. The only problem is that your potential customers don't know about it, and there's no easy way to discover it.
I noticed users in the United States can't sign up. ("United States" is not in the list of countries in the required Country field, and it's also excluded from "18. Regional Service" is the Terms of Service.) Is that intentional? It could be part of the reason for the slow uptake.
I’m not sure what you’re trying to say. I’ve tested it personally out on some real clunkers.
Maybe the animations on the landing page don’t play nice with every browser… Hard to know without more details, but I guess I appreciate you checking it out either way.
Yes, it's the animations. They're very intensive in general but especially the ones that AI-generated sites like to use to get around the basic attempts to block them. I was admittedly a bit hostile because the frequency, intensity, and hiddenness of animations is becoming an increasing aggravation across the web, and more and more frequently they're disregarding my OS/browser settings to disable animations.
It doesn't look like they're "trying to say" anything. They said it.
They opened the homepage and heard and/or felt their fans rev up, which didn't leave a good impression. They don't have confidence that your product/service is worth paying for (perhaps not worth using even if it were free).
People put that stuff up on Google maps, Facebook, and Instagram now.
I know it’s not popular with the crowd here, but those platforms are free, easy to use, and where the customers are. The mainstream options for a website like squarespace are absurdly expensive.
Yes and no. I find the restaurant on Google maps but 9/10 times the menu is either outdated or not properly structured and having a link to the menu website is better. So Google maps is the top of the funnel but I still appreciate a website.
I can’t help but think what this means is just that the menu isn’t that’s important as a marketing tool. If having an up to date website and menu resulted in a noticeable boost in business, every restaurant would have it.
Average person either finds the place through google maps or a TikTok video, checks a few photos of the food or venue, then goes. Doesn’t matter what the exact menu is because there are plenty of options and something will be appealing.
Or it’s good for customers and bad for restaurants. There are such things, and menu can be easily one. Especially tourist focused restaurants infested with such tactics, and you can avoid most of them just looking on their menu.
Yeah that context matters significantly. What’s the turnover rate for restaurants in your area? What’s the variance in menu? “Success” in my neck of the woods is staying open more than 2 years, and menu availability plays a significant role.
We usually order by phone, then drive by and pick up the food. Can't do that w/o a menu. The solution is usually to take a printed menu with you when you're there. But that's a chicken-and-egg problem!
I think it's important for customers and they usuallly post the menu in google maps thing, basically the customers are doing the labor of the business owner and the business owner as he still gets the results he doesn't do it
The conversational context did not involve anyone making any claims about the viability of businesses operating sans info. You can check—nowhere does the person who you're responding to (or any of the ancestor comments in this thread) write in their comment that companies are losing business because of the lack of up-to-date information, whether on their own site on through Google Maps.
The context is people, very reasonably, making a plea that that info be published on the open web.
I think the parent is making the assumption that a business owner would be able (and willing) to update the menu on their own website, whereas random pictures on Google Maps/Instagram might not have the most recent menu.
Google maps makes sense at least, but you're straight losing money if all you have is an instagram page. I can't tell if the facebook mention is a joke or not.
Menus change ie seasonal, and there is a daily changing handwritten chalkboard: Make a photo, put it on IG. Hours change: This week only opened from 8 instead 7: Post it on IG. Who has the time to answer a phonecall? And who uses phone numbers these days anyway? Text me on whatsapp like everyone else does.
Disclaimer: Don‘t use IG. But if I want to know if our favourite pizza place is open (cook travels to football games a lot), I ask my wife to check on Insta.
It's a trend in Sri Lanka for some reason to put your menu on Instagram... as a reel. Because you don't want your customers to have more than 15 seconds to view what you serve.
Not really. I don't have an IG account, but when picking a place ein an area I don't know, it is the place to get an impression of the place. The visual part tells a lot about the place, while many websites maybe got a photo from the outside, if at all.
How do you even know that a restaurant exists in an area you don't know? Not through Instagram, but through a web search or Google Maps.
But you're right, having a ton of good images on your website is one of the most important things for restaurants and most other businesses. And most fail in this aspect.
Google maps works the same way, thats the default in most of the world. I don't even know anyone who has IG account, myself including. Everybody has google account, not that you need one to browse (more or less) categorized photos on maps.
To be fair the Google maps restaurant side of the operation is quite possibly the largest ratio I've ever seen between "amount of capital and engineering skill available" and "quality (lack thereof) of UX." You have to access your restaurant profile through the Google search portal. It's a nightmare.
But in reality there’s only a handful of things people care about for your restaurant: what, when, and where. Put up your menu, put up your hours, and put up your location. And a phone number.
It's those things but more as questions than things they want to read. What people actually care about for a restaurant is:
"Can you tell me if the food is good?"
"Can you tell me are the staff great?"
"Can you tell me what does it cost?"
and "Can you tell me where it is?" to an extent, especially if it's not on a major route.
People want answers that they can trust for those things. They want a trusted source to tell them the answers.
You can't really get any of those things from a Google search or a website (ignoring reviews because they're gamed to hell now). The majority of a restaurant's customers come from word-of-mouth recommendations or reputation through curated services like critics and directories especially at the top end. A good website helps for people who are visiting the area, or for restaurants that are very new and whose owners don't have a great network (or who wrongly believe a website is key to getting business), but for most restaurants the only way to drive business is to build a loyal base of people who tell their friends and colleagues about it.
If a restaurant is going to have a website at all it should be a great one, because bad websites shouldn't be a thing, but a restaurant could happily run for decades with just an Instagram page these days and it'd make no difference to their success.
Sadly, at least in the Netherlands, most restaurant have to pay extortionary prices to aggregator sites like The Fork and others, that most people use to find restaurants and reserve a table. In addition they are incentivised to offer reduced prices on their meals, so the algorithm ranks them higher. So dominant is the role of the aggregator that the restaurant cannot afford not to be listed, and lose the customer base that flows in through these aggregators. Having their own website is of lower concern than doing this well.
Surveillance was always possible but it was expensive because a person or two had to literally watch a suspect. So it was rare in practice and suspects had to be chosen carefully.
Mass surveillance is not new in the sense that surveillance was always possible. What is new is the scope of power it gives to those who can use it.
reply