1,291 karma · joined May 28, 2024
This is blog post/comment section summary encountered many times per day.
The other side of it is people who seem to have 'gotten it' and can dispatch multiple agents to plan/execute/merge changes across a project and want to tell you how awesome their workflow is without actually showing any code.
kwargs = {'arg1':'val1', 'arg2':'val2'}
my_obj = MyDataclass(**kwargs)
Works fine for a dataclass with arg1 and arg2 attributes."missing x..."
pip install x
run script
"missing y..."
pip install y
> y not found
google y to find package name
pip install ypackage
> conflict with other package
realize I forgot a venv and have contaminated my system python
check pip help output to remember how to uninstall a package
clean up system python
create venv at cwd
start over
...
</end of time>
* treating red lights like stop signs
* completely ignoring flash reds
* stopping at flashing yellows
* ignoring school zones
* driving 10mph under the limit
* driving 20mph over the limit
* ignoring turn only lanes resulting in being dumped into oncoming traffic until it leisurely moves back into the correct lane
* waiting until the last minute to get into the correct lane to make a turn/take an exit
* ignoring signs indicating a lane ends and having to force its way into traffic
* ignoring signs a lane ends and changing into the lane only to have to change back immediately
* pulling out in front of cars, especially at night or other camera obscuring events
* ignoring cars merging onto the highway and blocking them until I take over and get out of the way
I treat it like cruise control. And based on the recent accident during which FSD decided to run itself off a perfectly straight, open, clear rode and into a tree, I keep my hand on the wheel.
Even in a geo-fenced environment, FSD is NOT going to work for unsupervised robo-taxi applications. Musk needs to get out of the way and let them stick LIDARs on these cars.
All that being said, it has also handled really complex driving in/around DC/Baltimore including some tricky merges/transitions that are hard to track as a human driver. It's knowledge of all of the surrounding vehicles allows it to make these maneuvers with far more confidence that I would. It also frees me up to track the bigger picture of traffic while not having to manage my speed/lane position. It's like driving with a friend who's not necessarily the best driver, but is basically competent. Let's just hope it doesn't decide to drive me in to a tree for no particular reason.
Is it not possible that new theories and breakthroughs could result from this so-called statistical pattern matching? The information necessary could be present in the training data and the relationship simply never before considered by a human.
We may not be on a path to AGI, but it seems premature to claim LLMs are fundamentally incapable of such contributions to knowledge.
In fact, it seems that these AI labs are leaning in such a direction. Keep producing better LLMs until the LLM can make contributions that drive the field forward.
Will finally implement that gravity in TTE, despite vowing not to. We all know how well developers keep promises.
Knowledge Growth
Will achieve enlightenment on the true meaning of 'enshittification', likely after attempting to watch a single YouTube video without Premium.
I found these actually funny. Cool project.
I haven't researched the topic of social media's effect on young people, but the common sentiment I encounter is that it's generally harmful, or at least capable of harm in a way that is difficult to isolate and manage as a parent.
The people closest to this issue, that is parents, school faculty, and those who study the psychology and health of children/teens, seem to be the most alarmed about the effects of social media.
If that's true, I can understand the need to, as a society, agree we would like to implement some barrier between kids/teens and the social media companies. How that is practically done seems to be the challenge. Clicking a box that say's, in effect, "I totally promise I am old enough." is completely useless for anything other than a thin legal shield.
On another note, and perhaps others are feeling similarly, but I am finding myself surprised at how little use I have for this stuff, LLMs included. If, ten years ago, you told me I would have access to tools like this, I'm sure I would have responded with a never ending stream of ideas and excitement. But now that they're here, I just sort of poke at it for a minute and carry on with my day.
Maybe it's the unreliability on all fronts, I don't know. I ask a lot of programming questions and appreciate some of the autocomplete in vscode, but I know I'm not anywhere close to taking full advantage of what these systems can do.
A simple example. I was coming out of a business driveway, turning left onto a two lane road. It was dark out with no nearby street lights. There was a car approaching from the left. FSD could see that a car was coming. However, from the view of a camera, it was just a ball of light. There was no reasonable way the camera could discern the distance given the brightness of the headlights. I suspected this was the case and was prepared to intervene, but left FSD on to see how it would respond. Predictably, it attempted to pull out in front of the car and risked a collision.
That kind of thing simply can not be allowed to happen with a truly autonomous vehicle and would never happen with lidar.
Hell, just this morning on my way to work FSD was going run a flashing red light. It's probably 95% accurate with flashing reds, but that needs to be 100%. That being said, my understanding is the current model being trained has better temporal understanding such that flashing lights will be more comprehensible to the system. We'll see.
As an aside, the fact that people pay for cable and still have 4-7 minute ad-breaks every 15 minutes make anything YouTube does pale in comparison.
I honestly can't believe how good is has gotten. It's not perfect and I maintain full awareness of what it's doing, but it performs incredibly well 95% of the time. It's really nice to feel like you're being driven by a mostly competent driver everywhere you go. Really, it makes about the same number (or less) of mistakes as other people I've driven with, but the mistakes are of a different sort, and after a while you develop an intuition of the situations that are going to challenge it.
Either way, I certainly won't buy another Tesla while Elon is running the place.
I very much don't understand the opposition to paying for some of the services we use every day. I get not wanted to pay 15-20/month for something like Apple or Disney, if you only want to watch a single show/film. But people act like you are crazy for paying for YouTube or Twitch, meanwhile they lose countless hours experiencing the frustration of advertisements. It's common to find discussions online where people are both completely opposed to paying for YouTube, with no real justification for that position, and highly upset at having to watch ads. Same goes for Twitch, which is arguably worse as the ads interrupt live broadcasts.
I use adblock everywhere else, but I thoroughly enjoy not having to worry about it on the platforms I use the most. It's not a scam to pay for something you use and enjoy every day.
https://www.youtube.com/playlist?list=PLoROMvodv4rOAvKVR_dyC...