I, individually, am certainly much more productive in my side projects when using AI assistance (mostly Claude and ChatGPT). I attribute this to two main factors:
First, and most important, I have actually started a number of projects that have only lived in my head historically. Instead of getting weighed down in “ugh I don’t want to write a PDF parser to ingest that data” or whatever, my attitude has become “well, why not see if an AI assistant can do this?” Getting that sort of initial momentum for a project is huge.
Secondly, AI assistants have helped me stretch outside of my comfort zone. I don’t know SwiftUI, but it’s easy enough to ask an AI assistant to put things together and see what happens.
Both these cases refer almost necessarily to domains I’m not an expert in. And I think that’s a bigger factor in side projects than in day jobs, since in your day job, it’s more expected that you are working in an area of expertise.
Perhaps an exception is when your day job is at a startup, where everyone ends up getting stretched into domains they aren’t experts in.
Anyways, my story is, of course, just another anecdote. But I do think the step function of “would never have started without AI assistance” is a really important part of the equation.
I really don't understand all the hyperbole around this bridge. It's a suspension bridge, so the relevant bits are at the pylons, which just happen to be on either side of a huge canyon. It clocks in at #14 on Wikipedia's list of longest suspension bridges, with a main span that is 603 meters shorter (2023 meters vs 1420) than the longest.
More interestingly, to me at least, is the fact that 31 of the longest 50 are all in China (as are all but two of the 24 in the "under construction or planned").
> The 2,890-metre-long structure, which took more than three years to complete, reduces travel time between the two sides of the canyon from two hours to two minutes.
Pretty impressive. I feel like things in the US take a lot longer and cost a lot more. The prime example is the second avenue subway extension which has been planned since 1920. But I just searched for a few significant bridges like the Gordie Howe bridge which took about 7 years and 6.4bn Canadian (connects US and Canada). And this bridge which seems a lot more of an engineering feat took 3 years and 8 months and cost between $280 to $292 million
Chinese civil engineers, and engineering orgs, are good because they get a lot of practice.
In the West, and especially in the US, individuals and orgs don't get practice, so when they finally do get a new contract they have to stumble around for 5-10 years figuring out all the institutional knowledge that was lost.
By the time they figure it out, the project is over budget, so it gets canceled, and then it's 20 years until the next half-hearted attempt. Lather rinse repeat.
At root, a lot of this stems from a "managerial" mindset in which people and skills can simply be "reallocated" on a dime. They can't. You can't uproot trees all the time. You plant one and then it grows over multiple human lifetimes.
To say nothing of the NIMBYism. To acquire the land for use, you have to fight some armies of lawyers retained by a population with a lot of disposable income. (Yes, the US for all her problems has the biggest pool of disposable income in the world)
I know it would be attacked politically, but I wish in the US we would be more open to hiring foreign firms for these kinds of projects. Could we have high-speed rail if we just asked some French or Japanese company to build it for us? And we should structure contracts with them in a way that keeps the plans from being changed for political reasons. "Sorry state senator, we can't alter the route to pass through that town without re-opening negotiations which might cost billions."
I think the lesson here might be to solve your problems on a more local scale, as the fit of the solution is as important as checking the boxes.
What works in Japan works in Japan because it was a solution built for the problems, politics, and people of Japan. Sometimes there are lessons we can take from projects in other countries, other times it turns out like Cali HSR where the proposed 'foreign' solution might be logical but not politically tenable.
The problem isn't physically building the rail, the problem is the legal framework in which governments operate in. Multiple rounds of environmental impact statements, eminent domain lawsuits, preferred contractor RFP's, zoning, permitting, endless red tape...
>> The 2,890-metre-long structure, which took more than three years to complete, reduces travel time between the two sides of the canyon from two hours to two minutes.
The construction timeline and travel improvement are comparable to the New River Gorge bridge, which was completed in the US in ~50yr ago back when systems were structured to and the people who ran them actually were capable of producing results.
Freakonomics just had an episode about how China is run by engineers, who get things done, while the US is run by lawyers, who prevent things from getting done.
This is the central thesis of Ezra Klein's "Abundance" book. Basically, things like this (or high speed rail, or public housing) don't get built in the USA because the government has hamstrung itself with so many rules and regulations that it becomes prohibitively expensive and/or tied up in lawsuits.
Places like China, for better or worse, are not burdened with the problem of making sure every constituency is accommodated.
As a bridge Carpenter, you are correct. I have never seen a construction project of that size for that price completed in that amount of time in the USA. Not even close. I've done projects that are worth over 20 billion today. I've done projects as small as a sidewalk repair. There's no way you can turn out that scope of work for that price at that scale in America it's not even close. They have to be lying about the budget there's no way.
> I really don't understand all the hyperbole around this bridge. It's a suspension bridge, so the relevant bits are at the pylons
I understand what you're saying, but the experience is quite different for the people driving over it compared to a bridge where it isn't a 2000 foot drop.
Is this true? Does the center of a canyon have higher wind speeds than the edges of that canyon? And what about gusts? I'd assume that boundaries are more turbulent.
I think the point being made is that if you followed two ridgelines that make up a valley up to a common summit you could just jam a plank in there. You've got the world's highest bridge. It's only 4 ft long, but it is technically a bridge.
I'd be more interested to know how they raised individual components into place. But I presume they just started with small cables, then used those to raise larger ones into place over time.
I understand the point being made, and it's a valid one. My point, though, is along the lines of "engineers/tech-minded folk often miss the bigger picture". Yes, from an engineering perspective, it doesn't matter a whole lot how far the bottom of the valley floor is from the bridge (though I'm sure it matters some during construction).
But the "user experience" of someone driving over the bridge is vastly different, to the point where I know specifically of some people who wouldn't be willing to drive over it, and it's not "hyperbole" to point out how high this bridge is compared to the ground below.
All I'm seeing is fairly straightforward fact-based announcements. "The tallest bridge has opened - here it is." If that doesn't interest you, fine... but the reports are not hyperbole.
This project is beautiful. This is an incredible work of art. It might not be the longest, but have you ever tried to pull cable over 2,000 ft hole? Have you ever seen what it takes to actually do those columns? The work looks nice very nice design. It fits with the landscape very well. And the fact that it cost only 140 million is an incredible. For a comparison if you look up one of the bridges I did. We spent 280 million on this
We spent 280 million on this
Years ago I remember reading about an economist who stated something like, "the best way to stimulate an economy is to pay people to dig holes and then fill them in." (I wish I remember who said that.)
In modern times, that translates to paying people to build roads and bridges. Why pay people to sit on their butts and eat bon-bons when you can pay them to get something of value?
In more tangible terms, building infrastructure does elevate peoples' situations.
I feel like the metric needs to be "greatest distance of road from solid ground" or "greatest distance from linear interpolation between ground attachment points".
Time zones are probably the biggest limiting factor, followed by remoteness. In my experience, it's really hard and pretty slow to onboard a remote worker if you haven't already worked with that person in the past. And at a startup, you don't usually have the luxury of time on your side.
> Remote work from Europe is harder in this regard, and from India... would be night shifts only.
I do not mind working all day until I pass out, and I do not mind adjusting my life to a different timezone as long as I get paid enough, and considering that USD > HUF, it is probably a no-brainer.
So yeah, hire me for full remote! Unfortunately traveling is out of the question due to disability.
> The worker gets to leave their sponsoring employer on the day of their arrival, if they choose to. The employment contract may not include any clawbacks of anything.
I'm not familiar with current H1B law, but what prevents this from happening today? I've hired away an H1B holder in the past; the process wasn't particularly difficult.
My understanding at the time was that the tricky thing for H1B holders is that they can only have a 60-day gap of unemployment before they need to leave the country (or find a different visa resolution, I guess).
Now, if this new fee applies to H1B transfers as well as the initial application, well, that'll actually make it harder for H1B holders to change jobs.
No, not really. The features that distinguish between feudalism and capitalism are mostly to do with class mobility and hierachies. Capitalism without class mobility is indistinguishable from feudalism. All it takes is a couple of generations of unchecked markets and power accumulates completely in upper classes and we are back to feudalism. You can argue "no, because reasons" (paraphrasing) all you want. Good luck convincing me, though.
In the industry, that’s known as face (or facial) detection, which is a different problem than face recognition.
Face recognition means computing which individual from some other database of people a particular face belongs to.
There’s also face tracking — detecting a face in an image and then tracking the same face across subsequent images. Which is often implemented by using a face recognition approach, but without any predefined catalog of people — you just dynamically fill up your face database as faces appear in the image sequence / video source.
There is so much fun to be had with a rotary encoder and some sensors. What are your plans for extensibility? I think this would be an awesome hacker device — everyone needs a thermostat anyways, and an easy way to build new screens (volume control, door buzzer, whatever) and deploy to the device would be so cool.
I expect you’ll need to be heads down on the hardware and basic software problems to hit your dates. But I also think it’d be worthwhile to figure out the baseline for extensibility early. Maybe this is just a call-home mechanism so you can advertise updates, so you can do something more in the future.
I also wonder if you could somehow take advantage of ESPHome here, for very basic HA etc integrations (of other functions, to be clear).
Also, what are your thermostat algorithm plans? Are you intending to consume HA thermometers / sensors? Or perhaps expose programmability hooks directly on the device?
What language are you writing in? I recently did some rotary encoder work for a round display on an ESP32, and found the dev kit micro python stuff to be terribly slow. I’ve had good luck with LVGL in C++, and my “ick” feelings about c++ are pretty much totally resolved by a healthy dose of AI chat bots.
Send me an email if you’re interested in more info.
Oh, good to know. I have been using Toit for all the business logic, but the display can have special handling for performance if necessary. Toit can also adapt into custom C++ code if needed. https://docs.toit.io/language/sdk/display
First, and most important, I have actually started a number of projects that have only lived in my head historically. Instead of getting weighed down in “ugh I don’t want to write a PDF parser to ingest that data” or whatever, my attitude has become “well, why not see if an AI assistant can do this?” Getting that sort of initial momentum for a project is huge.
Secondly, AI assistants have helped me stretch outside of my comfort zone. I don’t know SwiftUI, but it’s easy enough to ask an AI assistant to put things together and see what happens.
Both these cases refer almost necessarily to domains I’m not an expert in. And I think that’s a bigger factor in side projects than in day jobs, since in your day job, it’s more expected that you are working in an area of expertise.
Perhaps an exception is when your day job is at a startup, where everyone ends up getting stretched into domains they aren’t experts in.
Anyways, my story is, of course, just another anecdote. But I do think the step function of “would never have started without AI assistance” is a really important part of the equation.