bdamm

Personally I loved it. The Japanese aesthetic is subtle and delightful. Also, riding my bicycle in the countryside with very few traffic signs to mar the view. Just the necessary ones, at exactly the right places. They assume you are not an idiot, and somehow it all works out!

Ah, the fixie benefit is in every stroke, especially the ones where I'm slowing down with my legs and doing whatever else I want to be doing with my hands that doesn't involve the brakes. I wish I could transmit to you the pure zen-like joy that I experience when riding the fixie. It is not "worse case" anything. It is absolutely a wonderful experience and I will never, ever, give it up.

You do you. I have a fast and light road fixie, and I have a nice carbon fiber road bike. Do I ride my road bike in fondos? Yes. Do I ride my road bike around town and for training? Mostly, no.

Why?

Never in my entire life have I ridden a flawless bicycle. Inevitably, small imperfections in the drive train creep in. A flawless bicycle isn't flawless for more than a few hundred miles. All it takes is one whack on a bike rack, one hurried toss, one oops where the bike falls over as I'm grabbing my gels from the pantry. And never mind actual crashes! Ever broken a derailleur hanger?

The fixie keeps going basically no matter what. A problem for the fixie is the kind of problem that sends me to the hospital. Which is to say, that the fixie can take a lot more abuse than I can. Does the drivetrain wear out? Sure! And it costs basically nothing to replace!

Personally I think any serious cyclist should have one. They are super fun to ride, robust, cheap to work on, and they make you stronger. What's not to like?

Are you thinking of affirmative action?

Affirmative action is by design discriminatory, but not against nationality. It's discriminatory based on race and sex. So I think your grudge is not striking the right target. And in any case, affirmative action has been mostly wound down, which began to happen when Obama was President. Not because he did anything, but because SCOTUS declared that his election was evidence that affirmative action was no longer required and thus ruled against it in new cases.

Could it be a right-to-repair issue? That seems to be the only legal wrench available for forcing automakers to open up access to anything.

It is kind of a hard problem to solve. Sure, we can have expensive and custom equipment flying around, but the invasion of Ukraine (and really, any war when it gets bad enough) is all about economic efficiency. The question is not can it be done; the question is, can it be done repeatedly, can it be scaled, and can it be cheap?

We will have a fully autonomous Siri that is actually good before Ukraine gets autonomous motherships making targeting choices to repel Russian assault corps. However, given the pace of the war and the pace of AI development, that may actually happen. But it isn't happening today.

Ah hah, this is it. I was also confused - the tool isn't the thing. It's the behavior analysis capability.

If anything this video is kicking off a storm of amateur hacking enthusiasm for these wild bikes. Do us all a favor and sell yours!

Autonomous weapons are actually a really interesting branch of this and deserves a little more.

Yes, autonomous weapons were explored and were found to be poor performers compared to actual pilots. The breakthrough is in terminal guidance and dozens of other little techniques to get quality human control extended into the far reach of the battlefield. And of course, AI assistance in logistics and analysis. But actual autonomous weapons making any more of a choice beyond "something is moving, kill it" have been, at least for now, mostly a dead end.

This is because it's very difficult to economically load the rather sizable compute requirement into the compact one-use weapons, and of course reliable communications aren't assured either.

That will probably change some day, but for now, cheap automous command drones making battlefield analysis e.g. mapping out enemy movements from afar and launching cheap autonomous kamikaze drones is not a thing beyond occasional limited testing.

Check back in 5 years. This is going to age poorly.

The reason is that, despite what many think, AI actually is able to create novel ideas and solutions. That's why AlphaGo was so important; it couldn't beat the world's best Go player just be being a fancy autocomplete and a big processor. It had to create new discoveries and then use them effectively. That was the turning point. It's been a decade of improvements since then, and AI is already making discoveries we couldn't have made without it. The impacts are already here and in your world, you just haven't recognized them as such yet. But in a few years it will be undeniable to even the most uneducated observer, since changes that could not be possible will be present in every person's life as the effects ripple out across basically every industry.

Yes, although in all honesty the actual compilation is somewhat small compared to the many layers of unit, simulation, integration, and regression tests. All those corner cases take a long time to explore.

You are absolutely right. It will be a small window in the development business. Enjoy it if you can!

Since I've been working in C++ a lot recently I decided to watch the video as I waited for a build to complete. So the length is about right. And fortunately, the video is a delight!

This is absolutely the right vision imo.

Personally, I'm seeing massive improvements to my workflow and the quality of the product I'm shipping. I'm using AI to crank out far more tests than I used to be able to write, and I am using AI to analyze results with far more fidelity and speed than I could ever have done myself. That means I have more quality time.

But this will change, because the meaning of software development will change to expect, nay to require AI use. I've heard this is already happening at e.g. Google. The expectation of what can be achieved by tinkerers and by professionals will change. The expectation of what it means to interact with software via your own agents will change and will become commonplace. Apple still hasn't figured out the local agent on the iPhone, but they will. 2027 is not going to feel at all like 2025.

But is any of that a fundamental change? It sure feels fundamental to me, but maybe that's because my everyday has totally changed, but the product I am responsible for has not. Yet. The product I am responsible for operates in critical infrastructure where I personally hope AI never has deep roots, but maybe that's just me. I don't think using AI to build a system that is offline from any AI is the same as depending on an AI to make realtime decisions for critical infrastructure.

Is this supposed to work? I just get a vast blank canvas with decorations but no content.

The launch business is impressive relative to the launch industry, but in terms of company capital value, it is a very small part of the company.

I'm rooting for SpaceX, but even I can see that there's some extremely dicey work being done to mask the enormous sinkhole of XAi (and Twitter).

Unfortunately, that isn't good enough. Any government has a library of exploits in free software, and there are many documented cases of this. Your encrypted files are only as good as the clients that access them, and that's where the weaknesses lie. Along with the traditional aspects of surveillence, of course. "Amateurs hack systems; professionals hack people."

For the vast majority of people, the hassle simply isn't worth it, especially when the massive reduction in convenience is paired with basically no improvement in security.

What is your point? If anything, PRISM argues that Apple users are more secure; it took the NSA a solid five years longer to work their way into Apple. But at this point, every provider is compromised, every country is compromised, so again - what is your point? The government knows a lot about everyone, it has nothing to do with choice of computing platform.

It is also worth noting that the USA is not the only country collecting data on US persons (and everyone else.)

The question is not, does the USA know about you, the question is really, how many governments know all about you and how does that information manifest in your life?

What are you saying, that you think Windows or Linux are better platforms for evading covert government oversight?

Let's be totally clear here; if the government is interested in you, your choice of computer platform matters very little in terms of hiding information about your life.

How? My experience with Excel, Word, Powerpoint, event Teams, is that they generally work fine. This is unlike the situation from e.g. 20 years ago, when you could barely get work done due to all the crashes, but that is a very distant memory now. There was a brief time during 2019 when Teams on Mac was kind of awful, but that's long ago in the past as well.

My biggest complaint these days is that Teams uses far too much CPU when I'm sharing my screen. But other than that, everything seems to be ok.

Except, it was true before and it is still true today that the best "artists" whether graphic or mathematic, are the ones that do somehow manage to cross the chasm of pure research and providing a tangible benefit to their benefactors. That aspect of understanding your customer is not changed by the presence of AI.

Is this a known or quantifiable thing? I thought that the limit had already been determined i.e. the existing models top out and at some point it doesn't matter how much time or energy you let the model consume, it won't improve the result. And with regards to training parameters, I thought we were equally limited there, e.g. the existing models can't benefit from a larger parameter space.

I was under the impression that improvements are arriving via how the models are trained and how model prompting context is constructed, rather than just by how much data or how much energy is spent searching over the model space for a particular prompt.

Is there some evidence that we have not reached a pleateau with just resource consumption on existing models?

They can definitely recognize the problem class and build programs to do math. So what's the difference?

It's like saying that people can't turn high torque nuts on machine bolts, because you can't use your fingers to do it. But you can use a wrench, so effectively, we can turn high torque nuts on machine bolts even though it isn't something we can natively do unaided.

Even if the code was like that (it isn't), the power of the current crop of models to analyze data for patterns and build context out of code is leaps and bounds what it was even a year ago. And any developer will tell you that the hardest part of fixing a bug is knowing where the bug is in the first place. Once you know where it is, fixing it is usually trivial.

There is serious magic happening in the construction of model context.

It isn't necessarily the case.

Another wishful/hopeful thought is that the human experience itself is valuable, that competing for resources and living within a social network and having physical needs somehow creates value that is essential for companies to operate.

But is it really the case? I don't think we know that, and I don't know if the economy that results when all the white collar and much of the blue collar workers no longer understand how to participate in whatever the economy is becoming. Because it is starting to look like old money is coming around, and soon we will all be serfs to the creature comforts of those who have money now, upward mobility will be a thing of the past, and a small ruling elite over the vast subservient majority will form, reorganizing societies to more resemble middle ages lordship rather than what emerged in the 50's and 60's following WWII.

"Impact" as in an inevitable crash. I like it because it has unquestionably caused the avoidance of two cases where my car would have hit something. Once when I was merging onto a freeway and the car ahead of me basically brakechecked panic'd at the last second while I was briefly looking over my shoulder at the traffic I was merging into, and the other when a friend of mine borrowed my car and nearly hit a deer. Both cases the AEB kicked in, the car came to a very aggressive halt, and the crash was avoided. Yes, the AEB has kicked in at other times, but on the whole, it's been great and I appreciate having it. Probably other manufacturers have different implementations and different experiences. Mine is a Tesla.

You're misunderstanding how difficult it is to make major architectural changes to products the way Apple can. One of the ways to do it is to hide the architectural change as something else, something niche, and only when it has survived the fire of deployment there try to scale it up to the full market. It's actually quite genious, and you can expect more of it now that Apple's hardware guru is the chief.

I can't help but wonder if this agentic-via-accessibility angle is the result of this new leadership. If it is, it's a very good sign for Apple, because software and especially the AI gap is Apple's achillies right now.

Resolving basic usability issues shouldn't require infiltrating the company.

This is just fearmongering trope. You can imagine whatever you like, but there's no evidence that they're anything other than a car and technology company that wants to sell lots of its product.

My car has AEB and it's great. I'll never drive another car without it. Why not take the energy out of the impact? Humans aren't perfect, and even less so as we age.