The standard tax deduction for car travel is $0.70 / mile in the US, which accounts for things like insurance, gas, maintenance, and depreciation. So $500 / month is around 700 miles, which probably around 90% of US drivers surpass.
I used to feel this way. In the early days of "ride sharing," I preferred Lyft and would sit up front so I could have a conversation with the driver, which they encouraged. It was really fun for a while, and I enjoyed meeting people from different walks of life. Over time, though, transportation became much more functional for me, and now when I take non-autonomous rides, it's more irksome than enjoyable when drivers strike up conversations.
Why the change? I think a big part of your experience is the fact that you "rarely take taxis." Once you're doing it daily or near-daily, the amount of smalltalk becomes more tiresome. Also, with kids and a busy life, I'm usually either looking to get things done or enjoy a rare moment to myself as I'm moving from place-to-place. I agree with OP that Waymo is a huge step up on those dimensions. There's no other human in the same space to feel awkward around.
The fact that they drive more safely and smoothly is a huge improvement, as well. Ironically, I thought this was going to be something I would hate about Waymo. "You mean it drives the speed limit and follows all the traffic laws? It will take forever to get anywhere." It took approximately one ride for my perspective to completely flip. It's so much nicer to not feel the stress of a driver who is driving aggressively or jerking to a stop/start at every intersection. It's not like you can tell them to just ease up a bit, either. When we ride with our kids, we feel massively safer in Waymos.
Yes, it will be disruptive, and I don't particularly love the dominance that big tech has in all of our lives, but I do think Waymo is a marvel, and I hugely appreciate it as an option. As soon as they can take kids alone to all their various activities, it will be yet another massive unlock for parents.
Driving to work is the most common way of commuting everywhere in the US except NYC. So in that sense, no, taking a taxi to work daily is not normal, just as walking, biking, and taking public transit aren’t normal.
When I worked in San Francisco I took Caltrain to the city, but I took Waymo from the train station to the office. San Francisco, like almost all US cities, has poor local transit coverage. In my case there was a bus that took a similar route, but it only ran every 20 minutes even during commute hours and wasn’t coordinated with the train, so if everything was running on time it would have been a 17 minute wait (plus an extra 5 minutes walking). I was busy and well paid enough that spending the extra $10 to save ~20 minutes of travel (and the uncertainty of when the bus would arrive, and how strongly it would smell like piss) was well worth it.
According to [1] the median Bay Area big tech worker earns $272k/year - or $130/hour.
According to [2] Uber drivers make $15 to $25 an hour, before expenses like fuel.
So while it's not normal it's certainly plausible that some people take taxis on a daily basis.
More broadly, as levels of wealth inequality rise in a given society, more people end up working in the personal service sector doing things like cleaning, food delivery, taxi driving etc.
In April 2016, Rovi acquired TiVo for $1.1 billion.[8]
In December 2019, it was announced that TiVo would merge with Xperi Corporation. The merger completed in May 2020.[9]
Xperi itself also split apart in 2022, so it's effectively 3 companies removed from its original roots. Basically at this point it is only valuable for the vague nostalgia consumers have for the brand.
If everyone is using LLMs to write new code, and LLMs are trained on existing code from the internet, that creates an enormous barrier to the adoption of new programming languages, because no new code will be written in them, therefore LLMs will never learn to write the code. It is a self-reinforcing cycle.
I've experienced this to some degree already in using LLMs to write Zig code (ironically, for my own pet programming language). Because Zig is still evolving so quickly, often the code the LLM produces is wrong because it's based on examples targeting incompatible prior versions of the language. Alternatively, if you ask an LLM to try to write code for a more esoteric language (e.g., Faust), the results are generally pretty terrible.
Fine-tuning existing base models on your programming language is pretty practical. [1] You might need a very good and large dataset but that's hardly a problem for a programming language you're generating because you better have the ability generate programs for fuzzing your compiler anyway.
[1] There are a lot of models that achieve this. E.g. Goedel-Prover-V2-32B [2] is a model based off of Qwen3-32B and fine tuned on Lean proofs. It works extremely well. I personally tried further fine tuning this model on Agda and although my dataset was pretty sloppy and small, it was pretty successful. If you actually sit down and generate a large dataset with variety it's pretty reachable to fine tune it for any similar prog lang.
> enormous barrier to the adoption of new programming languages, because no new code will be written in them, therefore LLMs will never learn to write the code
Let’s see.
I’ve vibe-coded some apps with TypeScript and react, not knowing react at all, because I thought it’s the most exemplified framework online.
But I came to a point where my app was too buggy and diverged, and being unable to debug it, I refactored it to Vue, since I personally know it better.
My point is that just because there’s more training data, the quietly is not necessarily excellent; I ended up with a mixture of conflicting idioms seasoned react developers would have frowned upon.
Picking a less exemplified language and supplementing with more of your knowledge of the language might yield better results. E.g. while the AI can’t write better Rust on its own, I don’t mind contributing with Rust code myself more often.
> But I came to a point where my app was too buggy and diverged, and being unable to debug it, I refactored it to Vue, since I personally know it better.
One of the many pitfalls with using an llm to write code. It's very easy to find yourself with a codebase you know nothing about that you can't progress any further because it keeps breaking.
Let's not underestimate LLM's ability to do in-context learning. Perhaps it can just read the new lang's docs and apply what it already knows from other languages
But didn't LLMs read all the math books and can't really do arithmetics (they need special modes / hacks / python to do it I think)?
So why would they be able to "read" the docs and use that knowledge except up to pattern matching level. That's why I also assume, that tons of examples with results would do better than lang docs, but I haven't tested it yet.
While I don't like to argue for LLM competency, you have to remember that at the end of the day LLMs are word generators.
They will always be bad at math unless there is a major structural change.
So while they cant learn arithmetic they should be able to learn programming languages given that they are way closer to what it was designed and trained for.
What if we require LLM to write anything in Brainf**? If the language design is small enough to insert into our message every time, maybe it can work well.
People were saying we would all be getting in our cars and taking a nap on our morning commute. We are clearly still a pretty long ways off from self-driving being as ubiquitous as it was claimed it would be.
There are always extremists with absurd timelines on any topic! (Didn't people think we'd be on Mars in 2020?) But this one? In the right cities, plenty of people take a Waymo morning commute every day. I'd say self-driving cars have been pretty successful at meeting people's expectations — or maybe you and I are thinking of different people.
The expectation of a "self-driving car" is that you can get in it and take any trip that a human driver could take. The "in certain cities" is a huge caveat. If we accept that sort of geographical limitation, why not say that self-driving "cars" have been a thing since driverless metro systems started showing up in the 1980s?
And other people were a lot more moderate but still assumed we'd get self-driving soon, with caveats, and were bang on the money.
So it's not as ubiquitous as the most optimistic estimates suggested. We're still at a stage where the tech is sufficiently advanced that seeing them replace a large proportion of human taxi services now seems likely to have been reduced to a scaling / rollout problem rather than primarily a technology problem, and that's a gigantic leap.
The story of the bank built from bricks sent through the mail reminds me of the time I completed a move from Austin to Boston by packing all my possessions into rubber tubs and sending them by parcel post.
The delivery date was a range, and I wasn't there on the day of the first attempted delivery. When I called the post office about it, their response (in a thick Boston accent) was, "oh, so you're the tub guy, huh?"
All in all, it was a really convenient way to execute a cross-country move, assuming you don't have a lot of stuff!
Back in that brief window when Amazon was bribing USPS to deliver on Sundays and I could get 50-75lbs of bird seed for $12 shipped I had lots of fascinating Sunday mornings watching postal service workers swear at me and heave bags at my front door.
Or a TON of checked bags. Ran in to a guy in the airport once checking 10 bags. He bought the cheapest suitcase sets he could find, packed what he could, and sold the rest.
When I moved internationally, I found out about the ‘M-Bag’ service. The post office gives you real mail sacks (hefty, expensive seeming things!), which you can directly fill with books and printer matter (and nothing else!). They’re then tagged after sealing the drawstring, and shipped internationally!
I’m sure the USPS wants those sacks back, but the post office in the UK, where I had them sent, was just perplexed by them and told me to keep them.
I don't know about the author's case, but often asset purchase agreements will make the principals / shareholders party to the agreement personally with specific liability provisions. If there are no assets left in the company, the buyer has no recourse against it, since it is essentially an empty shell (in certain cases, insurance could be an exception to this). As a buyer, you will want to have some protection against issues you don't know about at the time of sale (perhaps because you weren't told about them, or the seller was negligent).
Some words in Shakespeare have different meanings today or have simply left standard usage. I don't think the presence of a word in Shakespeare means it is de facto good style to use today.
From a correctness stand-point, I think a descriptionist would be satisfied with an attested usage, especially from such a source. From a style point of view, I still find myself feeling embarrassed for the author when I encounter this usage (which is my own problem).
Maybe because JS files (specifically add-ons) run from the local filesystem are given escalated privileges compared to normal usage, perhaps for ease of development. I'm just speculating, though.
I think it’s a limitation on all extensions applied by Chrome/Firefox. My guess is to stop extensions from making you force install more extensions or something...
(Also what’s up Andrew! YC S09 represent :wave_emoji:)
