x187463

HN has a huge user base. If the article is pro-AI, you'll see top comments praising AI. If the article is anti-AI, you'll see top comments deriding AI. The article and headline is just an attractor to those whose opinion is aligned. It's not unique to AI, basically every thread has your typical nerd-fights, but AI has been particularly polarizing for many reasons.

Not only is this NOT the case, but all Tesla vehicles since something like 2021 have included matrix lights. They have adaptive beams to automatically darken sections of the headlight beam to avoid blinding other drivers.

What's 'allowed' in a speed run is decided by the community who do the speed running. Further, run leaderboards are broken into categories based around defined parameters. So, perhaps there would be a category unique to hardware manipulation such as this. For major speed running games, the rules are often very well defined and enforced.

On a Macbook, natural scrolling feels right on the touchpad. The real crime is not having a separate setting for a mouse. I had to use applescript tied to a keybind so I can use natural scrolling on the touchpad and toggle to regular scrolling when using a mouse wheel.

I wonder how a focus on per-token API profits will impact the incentives to improve token efficiency and drive down costs through optimized compute. I suppose as long as a few leading labs are competing, we'll see progress in this regard, but it's certainly less in their interest than it is with a flat subscription pricing model.

With iMessage, Discord, and Signal, I can't imagine a use case for me or anyone I know to use WhatsApp.

> Without guidance, LLMs tend to paint themselves into a corner, because they’re generating code to solve individual prompts, not thinking holistically about an application’s architecture.

I've found I can prevent the LLM, in many cases, from thrashing on a bug/feature for long periods of time by switching into plan mode and, even in the middle of a conversation, having it reassess the structure around the problem, first. If you keep prompting about the same bug, it may keep producing variations of the problem code. But forcing it to stop and 'think' for a bit, has yielded much better results.

The course has all the materials you need. The book itself is sufficient, but I think the course adds a bit more detail to the explanations.

I've been finding a lot of ~5 year old satellite and street view data. It's only anecdotal, but it seems like Google is not updating their imagery as often as they used to.

I must believe, as with the medical profession's uptake of AI tools, if these tools prove themselves to be reliable and meaningfully helpful, she will experience more than enough professional influence to learn the tools.

Measuring token usage as a productivity metric is like measuring keystrokes. Don't mind me, just over here rolling my face on the keyboard for an hour so I can take Friday off...

...except each keystroke has an associated cost, the sum of which may equal or exceed my salary.

+1 for NAND-to-Tetris. I combined it with a visual logic simulator so I could actually see the structures beyond the VHDL. I would love to go back and do Part 2.

Modern refreshed models use an eye tracking camera.

I'm at 96% FSD miles since they started tracking a few months ago. It works very well. It was unreliable until about 12-18 months ago, but it's been great since v13.

GM/Ford/etc. driver assistance systems are nowhere close to the latest FSD. It would be nice if there were other options, but there isn't. You can't buy a Waymo. FSD can go from parked in the driveway to parked in a parking spot at your destination without any human intervention.

No, you can start FSD from stopped using a button on the screen. It drives to the destination and parks itself.

Just needs to update the site to make it clear it's an independent port of the project. Then, modify the name to MacPad++ or something. Good to go.

My dude, go grocery shopping with a 2-year-old and see if you want them walking around. They'll be peeling a sticker off the floor for two minutes, then grabbing everything off the shelf. It's perfectly normal to cart the kid around so you can actually make progress through the aisles. They can reasonably follow you around between 3-4.

From a quick look, this doesn't have the model evaluate code quality, but it runs a heuristic analysis script over the code to determine the groan signal. Did I miss something? Why not leave it to the model to decide the quality of the code?

FSD controls every aspect of the vehicle's operation and explicitly demands human supervision. There's nothing misleading or incorrect in the name FSD (Supervised).

> You failed, in this case by hastening to cross the intersection as soon as the light came green, to account for the possibility of another driver's error. If you weren't taught to do that, as I was, then the mistake is not entirely your own. It was still a mistake, which you have already acknowledged would have led you into an accident had your vehicle not rescued you.

Even if we accept your interpretation of the situation as true, you're making the case for FSD. You can think of FSD (or other self-driving solutions) as raising the floor for bad drivers. If I'm a driver with some otherwise dangerous habits (nobody is perfect) then FSD is filling the gaps in my skill.

> Not for me. You're the one wearing power armor, remember.

But this is a joint interaction between the pedestrian and the vehicle. I can't make the pedestrian more aware. I can't give the pedestrian super-human reaction time. I can, however, give those traits to the vehicle. That's a major selling point for autonomous vehicles.

96% here, including DC and Baltimore. Besides the bizarre Navigation choices and waiting to long for lane changes, FSD has reached essentially zero interventions outside of bad mapping situations. I really wish Tesla would use better map data, for sure.

> Jumping a light in the dark?

I don't know what this means.

> Not thinking or learning to navigate by verbal instructions from your satnav or phone, instead of compromising the primary sense you must constantly use to drive without risking manslaughter?

Navigating involves reading street signs, block numbers, and traffic markings. These are all visual elements that can distract from safety monitoring. How many minor accidents result from driver's trying to figure out where they are, or need to go?

> I'm sorry, but if this is the standard, I really can't describe it other than it is...to say nothing of your considering safety less important, as you say, in the "inner city" that is my home.

My claim isn't that safety is less important in city driving, it's that driving is far safer due to lower speeds. There's more time to react and lower risk of catastrophic results when driving at 35mph. The challenge for a driver isn't sudden loss of control as you may experience at 65+mph. The city driving challenge is trying to track markings, signage, pedestrians, and parked cars while also navigating and managing the vehicle's basic operation. FSD can track all of that without distraction and leave the driver responsible for more human reasoning tasks.

It's worth noting that older Tesla's pre-2024, are stuck on an old version of FSD due to compute limitations. Recent FSD, generally, does not hang out in the left lane and is very good at recognizing when vehicles approach from the rear. It will move to the right lane to allow them to pass.

It's not a semantic issue, FSD is a completely different system, but many people mix up the terms when discussing these systems due to poor naming. Autopilot is just cruise control and lane keep. FSD handles navigation and full vehicle control. Articles discussing the dangers of Autopilot are making perfectly reasonable claims about a system which was poorly named/marketed, but they are not meaningfully relevant to conversations about FSD.

> It seems like it would be much safer for the autopilot to continue doing its best while sounding a loud alarm to make it clear that something dangerous is happening.

This is essentially what FSD does, today. When the system determines the driver needs to take over, it will sound an alert and display a take-over message without relinquishing control.

Sure, but the practical experience is that FSD is fairly predictable. It's just a matter of personal preference that comes from experience. I wouldn't impose a system like FSD on everybody.

One of my major complaints about FSD is the 'speed profiles'. You used to be able to set a target speed directly. Now, you can only select a profile. You're either going the exact speed limit, 2-3mph over, or essentially 'with the flow of traffic' which can lead to speeding +15 over the limit.

Here's a simple example from last week. FSD was in control on my way to work, stopped at a red light early in the morning before the sun was up. The light turns green and FSD does not accelerate. I figured it was somehow confused and I was starting to move toward hitting the accelerator myself when a car comes flying through the red light from the driver's side. I hadn't noticed this car, but FSD saw it and recognized it wasn't slowing down. I could see there were headlights, but it wasn't clear how fast it was going.

It's just nice having a 'second set of eyes' in a sense. It's also very useful when driving in unfamiliar cities where much of my attention would be spent on navigation and trying to recognize markings/signs/light positions that are atypical. FSD handles the minutia of basic vehicle operation so I can focus on higher level decisions. Generally, at inner-city speeds, safety and time-to-act are less of an issue and it just becomes a matter of splitting attention between pedestrians, obstacles, navigation, etc. FSD if very helpful in these situations.

"Autopilot" is not $8K, that's FSD. Autopilot was the default cruise control/lane keep software and was renamed "Traffic Aware Cruise Control" a few months ago. The original name was ridiculously misleading.