I think it’s a bit of both. It’s very important to struggle with a problem yourself. You need your brain to be thinking about it, turning it over, at least for a bit.
But when you’re new on a topic, it can take hours to nail the solution by yourself. I’m thinking of science & math practice problems. If you skip ahead to look at the answer, you can use it to work backwards, to understand where you were struggling and going wrong.
Crucially, you’re not just copy/pasting the answer. But taking this “shortcut” early can definitely be helpful. It’s very easy to spend a lot of time going down the wrong path. If you can take the shortcut when you sense that’s happening, and really try to understand why that wrong path isn’t right, I think the shortcut can help speed you up.
I think the same applies with LLMs. It’s important to not just have the LLM do it all from the start. But when you start getting stuck, having it give you recommendations is a good way to get tutored, so long as you’re really thinking about it.
That’s exactly what this is though, it’s giving you the precise “incantation” you need based on what you’re trying to do. Like “list 20 biggest files in current directory”
well said that’s exactly the spirit of greppers — not trying to guess like an llm, just surfacing the ready-to-run incantations people keep re-googling.
huh? "list 20 biggest files in the current directory" is a search query, or a llm prompt input, it's not any kind of precise incantation.. ?
I mean if greppers is just translating queries like that to related command-calls, that's fine but bizarre and nothing that would require a community-contribution-based database of mappings, that's just a search engine and/or llm mapping problem...
Just ask yourself if Tesla cannot be autonomous in their own merely 2 mile long tunnel with only trained teslas drivers how can you believe any other promises ?
Setting a lofty goal and failing is not to be rewarded or congratulated, or in this case compensated .
There are other stuff to do in the world both as far as technological advancement and leisure that could fill that time and man hours and produce a tangible ROI as opposed to "we are failing but at least we are trying"
By the same token the next goal set by Musk would be creating a wormhole if he hasn't said it already considering the amount of drugs that he takes.
Many people conflate the reasoning that is prevalent in SV for founders where they say "there is no price for failure". That is maybe true for the individual not when trillions of man hours and trillions of dollars are allocated.
By your world view no one should do anything because they could fail. Ycombinator shouldn't exist because most of them fail.
They haven't hit your personal standard. Who cares? Investors are happy with their progress and at the end of the day that is more important than people like yourself who probably do t have skin in the game
> > By your world view no one should do anything because they could fail
It is one thing to fail independently say at your own startup, complete different thing is having choose between keeping your job or do something that you feel like it's a dead end or it is deceiving the public even tho you are not convinced just because the higher ups ordered so.
THat's how trillions of dollars and most importantly trillions of man hours are set on fire
When they actually lose money let me know. Right now they're still making billions in profit. If that's your idea of success you must be making tens of billions yourself
When you announce goals and fail them is like when boxers or UFC fighters announcing they'd destroy an opponent or sail easily to win the belt.
When they lose of course people are gonna call them out regardless of the monetary purse that they were awarded for losing the fight.
And that is the way it should be considering you are dangling in front of people a huge improvement in their quality of life and then constantly failing to ship it.
For all its flaws Microsoft shipped Windows 95 after talking about it for 10 months before the launch. Not 10 years. And that was really a big quality of life improvement for basically billions of people.
Finally people don't care about what the marketcap of Tesla is or what Musk net worth is, those discussions happen among the fanboys and those who have false idols. People care about how a company can improve their life and despite all the fanfare, Tesla Motors have produced very little improvment through its history, and thye are not a startup either considering they are 2002-2025
Windows 95 is comically easy compared to most of the stuff that Tesla does.
|People care about how a company can improve their life and despite all the fanfare, Tesla Motors have produced very little improvment through its history,
This is just objectively not true. They produce some of the safest cars in the world, and have the best selling car in the world. Obviously those same people you are talking about disagree with you.
Tesla was founded in 2003, not 2002. Regardless they are still a startup compared to all the other US car manufacturers. Regardless though Tesla has done amazingly well, been very profitable while others are suffering, and despite major cap-ex are able to still make a profit.
> > Windows 95 is comically easy compared to most of the stuff that Tesla does.
Many would not agree but whatever , it was a paradigm shift that changed the world and improved the quality of life of billions of people.
The point of companies is to improve the quality of life of people, not pursue stuff that is hard for virtuoso technical bragging rights. That is something that theoretical physicists do. Actually if something is a low hanging fruit that can improve the quality of life of billions of people that should be pursured not the virtuoso technical mastership for the sake of technical virtuosity
> > This is just objectively not true. They produce some of the safest cars in the world, and have the best selling car in the world. Obviously those same people you are talking about disagree with you.
Put a car from 2003 and a Tesla from 2025 next to each other and they are 99.999% the same. As a matter of fact Tesla can't do many things that a 2003 car can do.
This is the folly of people who think that a product which is 200 year old like the automobile can be revolutionized just because you can convince a bunch of loonies that it became a "tech company"
And you are fixating on technical prowness for technical prowness' sake.
If a start button can improve the life of billions then it should be pursued.
There are many thousands of "start button" alike solution which would improve the life of billions but the capital and the man hours are tied into fixations which are impossible or perpetually 20 years in the future.
Between Waymo and Comma.ai, yes they did. Not in those goofy tunnels, but the underlying tech is there. So then why are the Vegas tunnels not self-driving? Tesla FSD works great in select places right now! How is the tunnel not one of them?
Not at all. It's traffic that is completely below ground which means that no streets above are burdened with it, it's all electric so no emissions. Name me any other form of transportation that was built so affordably in the US and moves this many people while not impacting street traffic?
By comparison the Vegas Monorail cost 5x as much, took 1/3rd longer to build, and takes roughly a third of the passengers that the Boring tunnel does. This is a huge win for Las Vegas, and they continue to build out because it's worked out well for them.
> It's traffic that is completely below ground which means that no streets above are burdened with it, it's all electric so no emissions
So are modern subways. Cost is a major point tho, subways are designed to move waaaaay more than 30k people a day for much less, but costs of building are much higher.
This is only 1.7 miles and a novelty, I would not know If the differences hold for Tesla on other places or when scaling up. My suspicious is that it does not.
I also wonder that if you use the same tunnel they did but modify the cars to run by themselves using traditional techniques, would the operation get cheaper but the shortcomings be more glaring.
Vegas now has 5 stations and is 2.2 miles. Can you realistically compare it to a billion dollar a mile sub at line that would take a decade to build (or more)?
30k a day is a nearly a million a month and costs are low by comparison (no expensive subway cars etc).
Have you actually looked what some of their extensions look like?
Please go look at Encore station on Vegas Loop. And come back telling me it is below ground... And that is not only place. Overall it looks like right mess. Including very scalable solutions like single tunnels to some stations...
Exactly. Before I was diagnosed with ADHD, I was experiencing severe executive function. Just nearly no ability whatsoever to do things I wanted to do, even if they were fun. But had big impacts in terms of friendships (not actively trying to socialize very much, because planning basic get together felt insurmountable) and home life (simple chores were almost impossible to get done). And it doesn’t feel like laziness (which I think is more like you just don’t want to do the annoying thing right now), but like I really wanted and needed to do it, and just couldn’t make it happen.
ADHD meds have been almost a literal lifesaver, and the amount they make work halfway tolerable is… crucial.
Passkeys is not security theatre, and also not a UX pain if you use a password manager. Turns out it’s nice to have a standardized API for submitting a credential to a website rather than relying on browser extensions to hopefully guess the input field is for a password. (Not to mention the multitude of sites that don’t properly handle text being autofilled)
There are exactly three nice things about passkeys.
1. It forces the use of keys with a reasonable amount of entropy, and the use of a password manager to access them.
2. They will not make it easy to use a key with the wrong site (also true of a good password manager).
3. Uses public/private keypair so key itself is never sent over the wire (even encrypted).
The real question is whether these properties are worth all the costs (enumerated in this article).
Agreed, my wife even says 1Password is one of the best tech things I’ve set up just because it completely solves sharing passwords and stuff with each other.
That basically does exist, and it’s called SSO. SSO providers (eg Okta) have a unified dashboard where you can control who can access what, and at what level, and can revoke access any time. It’d be nice if there was a version of that for families that wasn’t insanely expensive.
Anyways, 1Password completely solves this problem for me with me & my wife.
It doesn’t do anything for pushing authentication remotely or controlling access within apps, such as voice chat in Roblox. Each app has proprietary controls.
It also doesn’t begin to cover notifications. For some reason most services seem to think only one parent is in charge and both don’t need equal access and equal notice.
You hit the nail on the head too. Coding itself is very easy for anyone halfway decent in this career — and yet there were a ton of people in CS101 and even in later courses who struggled with things like for loops. It was very hard for them to succeed in this career.
What’s hard is coming up with the algorithm/system design, making the right choices that will scale and won’t become a maintenance nightmare, etc. And yeah, after almost a decade, I have picked up enough I can at least write an outline of a solution that will work alright. But there are still so many tricky edge cases and scaling problems that make it hard to turn “alright” into “really good!”
Sure, AI can help… but it mostly helps with greenfield projects. It doesn’t know about the conversations on slack & jira from a year ago. It doesn’t know about the dozens of other systems and ways the project interacts with other parts of the business. It doesn’t know why whatever regurgitated approach won’t be a good fit for our specific use case. And elaborating all of that detail is not easy! Part of what makes you a good employee is the shit you picked up over the past several months & years that is joe ingrained in your mind when you start working on new projects.
But when you’re new on a topic, it can take hours to nail the solution by yourself. I’m thinking of science & math practice problems. If you skip ahead to look at the answer, you can use it to work backwards, to understand where you were struggling and going wrong.
Crucially, you’re not just copy/pasting the answer. But taking this “shortcut” early can definitely be helpful. It’s very easy to spend a lot of time going down the wrong path. If you can take the shortcut when you sense that’s happening, and really try to understand why that wrong path isn’t right, I think the shortcut can help speed you up.
I think the same applies with LLMs. It’s important to not just have the LLM do it all from the start. But when you start getting stuck, having it give you recommendations is a good way to get tutored, so long as you’re really thinking about it.
reply