birken

I'd really like to see a regular poll on HN that keeps track of which AI coding agents are the most popular among this community, like the TIOBE Index for programming languages.

Hard to keep up with all the changes and it would be nice to see a high level view of what people are using and how that might be shifting over time.

What kind of headcount do you estimate $1MM/year can reliably support?

That's like ~2 engineers at FAANG.

It is difficult to say this is what consumers want, when right now consumers are getting the best of both worlds: The ease of AI agents without the long-term negative consequences of destroying the publishers who created all the high quality training data in the first place.

I think in the long term the highest quality content creators are going to find ways to keep their information out of AI training data, and put it behind walled gardens.

While yes, I am attempting to collect ad revenue from users, and yes, I don't want somebody competing with me and cutting me out the loop, a large part of it is controlling my content. I'm not arguing whether the AI chatbot has the legal right to access the page, I'm not a legal scholar. What I'm saying is that the leading search engines also have the equal rights to access whatever content they want, and yet they all give webmasters the following tools:

- Ability to prevent their crawlers from accessing URLs via robots.txt

- Ability to prevent a page from being indexed on the internet (noindex tag)

- Ability to remove existing pages that you don't want indexed (webmaster tools)

- Ability to remove an entire domain from the search engine (webmaster tools)

It is really impolite for the AI chatbots to go around and flout all these existing conventions because they know that webmasters would restrict their access because it's much less beneficial than it is for existing search engines.

In the long run, all this is going to lead to is more anti-bot countermeasures, more content behind logins (which can have legally binding anti-AI access restrictions) and less new original content. The victim will be all humans who aren't using a chatbot to slightly benefit the ones who are.

And again, I'm not suggesting that AI chatbots should not be allowed to load webpages, just that webmasters should be able to opt out of it.

The AI isn't "reading the web" though, they are reading the top hits on the search results, and are free-riding on the access that Google/Bing gets in order to provide actual user traffic to their sites. Many webmasters specifically opt their pages out of being in the search results (via robots.txt and/or "noindex" directives) when they believe the cost/benefit of the bot traffic isn't worth the user traffic they may get from being in the search results.

One of my websites that gets a decent amount of traffic has pretty close to a 1-1 ratio of Googlebot accesses compared to real user traffic referred from Google. As a webmaster I'm happy with this and continue to allow Google to access the site.

If ChatGPT is giving my website a ratio of 100 bot accesses (or more) compared to 1 actual user sent to my site, I very much should have to right to decline their access.

A datacenter could consume a lot of water with evaporative cooling. I don't know how prevalent it is, but given how cheap and efficient evaporative cooling is, I'd guess datacenters use it a lot where possible (probably in combination with other cooling methods).

Coincidentally I just had a professionally done garage door spring replacement today, and I asked the repairman this question, and here is what he said:

1. The springs lift the door from the bottom, and from each side, which puts less load on the door itself as compared to if the entire weight were being lifted from the top middle every time.

2. The motors can be smaller, quieter and use less power

3. In case of power failure, the door is much more functional and safer the less apparently weight it has.

Also the springs themselves are very unlikely to be dangerous (as long as you don't try to replace them yourself), because he said they almost always break when the door is at the closed state, because that is when they are under the most tension. Therefore on the whole, the springs in practice offer no practical safety risk, while greatly increasing the safety of the door in it's normal operation while also reducing wear and tear on the door. They also allow people to have heavier types of doors if they want them.

The people who come up with 409a prices have every incentive to make it as low as possible provided it is somewhat defensible to the IRS.

I assure you they can get more creative than saying that the last preferred price was at $X, therefore our hands are tied and the common must be close to that. They can take into consideration the preferred preferences, the current state of the business, the time since the last round, etc. For example, the 409a value can keep going down and down if the value of the business is (defensibly) going down and down, regardless of the last fundraising round.

And then they'd have sold at the higher value, because the sale price is almost certainly higher than whatever the 409a price was.

It might not be a 90% discount but it still will be a >50% discount

And how did that go?

Self driving electric cars are much better than car shares as they'll have significantly higher usage rates and eliminate much of the need for parking spots in dense downtown areas.

If you believe car shares are part of an efficient transportation future then self-driving cars are part of it!

I'd just like to note the irony of saying how bad Google is while at the same time saying that Google's general purpose search engine is still better than any of the specific search engines of these individual sites.

Perhaps the issue isn't that Google is bad, and the issue is that search is incredibly hard.

Generally speaking the people who live in SF aren't comparing SF to Boise, they are comparing SF to other elite cities like NYC, LA, Boston, Austin, etc. And the COL is not 2-3x higher in SF than those other cities.

Also as a little story about how crime infested SF is, a couple years ago my wife accidentally left her key ring in the lock on our front door (which included car keys to a car right out front), where they stayed for the next ~18 hours, right in obvious public view from the street, until we noticed the next morning. And lo and behold the crime was so bad that the keys were right there the next morning and we had a nice laugh about it.

You always need somewhere to live, so you can only "sell high" if that means you can "buy low" somewhere else. If all bay area houses go up 10%, then you haven't actually gained anything assuming you still want to live in the bay area.

You also lose at least 5-6% of the value of real estate every time you buy or sell. So given that buying/selling real estate is complicated and expensive, and moving is complicated and expensive, for the vast majority of people, the value of the house going up and down _in the short term_ is mostly irrelevant.

I've logged a lot of miles of Comma in a 2023 Hyundai Ioniq and 2022 Toyota Prius Prime, and the built in driving assist of both cars is nowhere near Comma, both in terms of steering and accelerator/brake.

Things that Comma handles seamlessly that the built-in cruise in both cars will not:

- Full stop and go

- Sharp turns on the highway that require slowing down (both built-in adaptive cruise modes will gladly just drive you off a cliff at 65 mph)

- Situations where the lane lines are hard to see or are implied

- Non-highway driving

- Not requiring me to touch the steering wheel every 20 seconds

Maybe those things work in higher end cars (though I'd say the Ioniq is a fairly high-end car), but then again with Comma you get it for ~$2k in a ton of cars instead of having to buy a luxury car.

It is true that if you are on a highway, with clear lane lines, the steering assist in both cars is certainly a lot better than nothing, but it's just not nearly matching the reliability and versatility of Comma in any sort of imperfect situation.

As a very happy user of Comma, I think it is reasonable to say the company is going to fail, but that ignores that the product they created is still awesome. Comma is light years better than any built-in driving assist in any non-Tesla car. And it's comparable to Tesla for far less money.

The reason the company might fail is because their main thesis, being that car manufacturers would just license the self driving tech to somebody else (like Comma), never came about. Car manufacturers are just too conservative. It was a perfectly reasonable bet to make though. Unfortunately they ended up in the business of selling hardware and giving away software for free when they wanted to be in the business of selling software.

Well I guess we'll both agree to disagree about the nuances of the early years of Reddit, something neither of us know anything about. Oh well.

There is a very big difference between giving no credit and giving a fair amount of credit, which might be small. If you listened to the podcast they talked about Aaron in a non-disparaging way, and a way that I'm sure is accurate to their beliefs about it.

But my point is that _even if_ Aaron was super, super important to the company 16 years ago (which I can't speak to), Reddit was a tiny little fledgling of what it is now. From what little I know about Reddit's history, I know that Aaron stopped working on Reddit very soon after he started and has had no contribution to the company since then.

I also do think it's interesting that the people that have worked at Reddit for years/decades all seem to have a similar opinion of the situation, and the people who haven't worked at Reddit at all are the ones that feel so strongly that Aaron is being wronged somehow.

Maybe Aaron himself was more proud of all the other things he accomplished! Saying he was an important but small part of Reddit's story is still something!

If I had to guess as to why his contributions are marginalized, it would be that his contributions were ~16 years ago to a version of the website that is vastly different in scope and scale to the one that exists today. Founders are important entities for companies, but just having the title of "founder" doesn't also magically mean whatever contributions and decisions you made have to be memorialized and defended forever after so many things have changed.

I have a very CPU intensive single threaded program that runs about 15% faster on the intel CPU. Other CPU benchmarks out there show my particular processor is slightly ahead of the M2 (and there are faster intel/amd processors than the one in my machine). But if you look at power consumption I believe the intel CPU uses something like 3-5x the power to generate that slightly better performance.

At the high end of performance, the M2 is just light years better than any other chip (other than the M1) in terms of performance per watt. But if that stops being a concern to you then I think most linux desktop users are better off just getting a "normal" intel machine, which also will have 100% driver support for everything, cost less, and probably be faster.

I have a relatively mid-to-top tier linux desktop (i5-13600K, 64 GB ram) and an M2 Macbook pro (M2 max, 32GB ram). The linux desktop comfortably outperforms the M2 in any CPU based benchmarks while also being significantly cheaper (though the M2 desktop machines are cheaper than the laptops, but still more expensive than the alternatives). If I were doing GPU heavy work than I assume the M2 would have a leg up, but then again I could just buy whatever GPU I wanted for the desktop machine and put it in there and I think it would outperform the mac once again.

However, if you were concerned about power consumption than the M2 would win by a mile, but for me it wasn't a huge factor in a desktop machine.

I got a Moonlander about 3 months ago, and it took me about a month of getting used to the key layout until I was typing at similar speed to a Macbook keyboard (which for me is fairly fast, about 120wpm). The hardest letters for me was shifting from my index finger to my ring finger for 'c', and shifting '.' from my other index finger to my other ring finger. I basically got used to never typing a key with my ring finger below the home row and with the ortholinear keys it was just no longer possible to type quickly while having my index finger handle everything. The keys also poke out quite a bit more than other keyboards so when you stretch to type a key with the "wrong" finger you are also much more likely to accidentally press another key.

Getting used to spacebar, backspace and enter, all of which I shifted to thumb mappings, took only a few days. You hit those keys so often you learn the muscle memory very quickly. This to me is by far the best part of the moonlander, having 4 super-accessible thumb keys and 2 semi-accessible thumb keys that I can program is a massive, massive improvement over a standard keyboard. Your thumb is completely wasted in a normal layout.

I still am far below my "normal" speed for programming, but that is because I'm trying to figure out the best custom mapping for (){}[]<>=+ and then train the muscle memory.

Overall I'm very happy with the purchase, I feel like it is much more ergonomic and with another few months of service and perfecting my layout it will hopefully end up being much faster and more convenient than a normal keyboard layout as well.

I don't know if all FAANG companies are the same, but your description of the job does not match my FAANG experience at all.

It was dehumanizing in its own way because you are such a small cog in a large machine, but nobody worked long hours, nobody was forced to be available all the time, nobody really needed to sacrifice for the company (because everybody was such a small cog) and even the lowest performers I worked with could basically not do anything to get fired. Many people there were also "lifers" because the job was so easy.

Most people left due to having ambitions for more impact, not due to being burnt out.

Changing all images to be lazy loaded is not a good idea for SEO reasons (and perhaps not a good idea for regular usage of the site either). If it were, browsers would just do it automatically. Using hints on your images is only useful if you are doing it strategically, which means only lazy loading images that are offscreen on the initial page load. Otherwise you are not actually giving useful hints to the browser on how to load your page in the correct manner, and thus you are just better off letting the browser use whatever internal logic it has to decide how and when to load the assets.

Basically you can think about the optimal, minimal set of resources to render your page being:

- HTML

- Required CSS

- Above the fold images

Then at that point, download all of the other things to make the page work in the way you want (javascript, etc). Anything else is delaying the initial page-load. Because your images are lazy loaded, your page load looks like the following:

- HTML

- Required CSS

- Some above the fold images (profile.jpg, etc)

- Massive, >300kb blobs of javascript (runtime-main.js, main.js)

- Above the fold images in the post

This is not good. It doesn't make sense for your page to download 300kb of javascript before downloading the 18kb image that is above the fold. Now you can partially solve this problem by making the javascript asynchronous, but that still is just another band-aid on the problem, as then the javascript and above the fold images would download concurrently, which is still not optimal.

What you want to do is have above the fold post images be loaded eagerly (the default), and then lazy load ones that are lower on the page. If you aren't going to do that, you probably are better off just not having the images being lazy loaded at all, especially if your page includes 300kb of javascript which is likely going to be much larger than the combined size of all the images on the page.

To be fair most people who are driving cars aren't towing caravans, so this seems like an extreme edge case.

I heavily use adaptive cruise control and I've never had any issues with it, though I'm always ready to take over in case anything complicated starts happening. It's a nice feature because it saves a lot of energy when the driving is very easy, and then you just drive normally otherwise.

I've had a little robot 3 for 5 years. It isn't the most reliable device in the world, but you learn very quickly all the little quirks of how to fix it when it breaks and it isn't too bad. I'm very, very happy with it. If you aren't handy then it might be more annoying for you.

My only complaints seemingly are fixed by the litter robot 4: better sensors (so it breaks less) and a larger poop storage. I would buy one immediately if I didn't already own the LR3.

Additionally, the worst strains on the grid are really measured by the peak loads, so if you change billing a way that makes the peak energy less expensive, and therefore disincentivize people to (a) reduce usage at peak times or (b) install solar panels/batteries to mitigate their peak usage, then it actually might make the grid more unreliable.

Water is different in that we really only have usage concerns about water, AFAIK there is never any issue with the water pressure drops at peak times (perhaps thanks to water towers). With electricity the concerns are mostly about the peak usage, and not so much the overall usage.