As a simple example: you know after signup 20% of customers purchase, 80% don't, but what you want to trivially add in is the fact that of the users in a marketing campaign, 10% of them signed up, which means for the marketing funnel 2% purchase. Now consider that you have 20 or more such events in your funnel and you want to combine them all with out doing all the math by hand. Likewise you want to be able to add a newly discovered step in the funnel at will.

Using a topological sort you can take an arbitrary collection of nodes where each node only knows that probability the next nodes are, sort them and then compute the conditional probabilities for any user reaching a specific node fairly trivially, given the assumption that your funnel does represent a DAG.

If you don't perform the topological sort then you can't know you have calculated all the conditional probabilities for the upstream nodes, which makes the computation much more complicated. Topological sort is very useful any time you have an implied DAG and you don't want to have to worry about manually connecting the nodes in that DAG.

Curious, did you seriously not recognize that this is a famous quote from the final scene in Blader Runner?

I read that more as quirky call out to a famous film, not that he was claiming this was his own view of the world.

I'm always surprised how useless something is when I don't know it, and suddenly once I do know it, I solve lots of problems with it!

I've heard programmers grumble about how useless calculus is, before I learned calc I used to grumble about that too. After I learned it there were countless problems I unlocked solutions for by applying the thinking I learned in calculus.

I've heard programmers say that you'll never need to implement your own sort for mundane tasks, but, it turns out that after really grokking topological sort I used it countless times for fairly mundane problems like creating plots.

I've heard programmers say that learning the lambda calculus is a waste of time, and nobody uses functional programming. Yet it was people that understood these things that transformed Javascript from a useless browser oddity into one of the most widely used languages. It was seeing that Javascript was essentially a Scheme that unlocked it's true potential.

Over my career it's remarkable how many "esoteric problems" have lead to me solving hard tasks or even shipping entirely new products. If you're only focused on what is required of your day job today you're only going to be at best a mediocre engineer.

For me, it's been the opposite: the last 2 years have been the best time I've had working in tech since the early 2010s.

Around 2019 I was seriously considering leaving the field (if it didn't pay so much) as the entire industry had turned into a bunch of leet code grinding, TC chasing, mediocre drones. It was incredibly hard to find people working on actual problems let alone challenging/interesting ones. Nobody I worked with for years cared one bit about programming or computer science. Nobody learned anything for fun, nobody hacked on personal projects during the weekend, and if they were interest in their field it was only so they could add a few more bullet points to their resume.

But the last two years I've worked with several teams doing really cool work, found teams that are entirely made up of scrappy, smart people. Starting building projects using a range of new tricks and techniques (mostly around AI).

Right now there are so many small teams working on hard problems getting funding. So many interesting, talented and down right weird programmers are being sought after again. People who like to create things and solve problems are the ones getting work again (my experience was these people were just labeled as trouble makers before).

I'm probably getting, inflation adjusted, paid the least that I have in a long time, but finally work is enjoyable again. I get to hack on things with other people who are obsessed with hacking on things.

If you're using Python, Jupyter still offers a major advantage for sharing prototyping/research work in that it it is the only real way to provide literate programming. There are cases where code inside of a notebook should be pulled into a library, but pure code will always be inferior when most of your work is expository and exploratory.

Literate Haskell, and R Markdown are superior forms of literate programming, but Jupyter is not terrible if you just take some basic precautions before sharing.

Notebooks work great as... notebooks. Which is primarily how they're used in my experience. When the bulk of your work is prototyping and and exploring an idea, they're still an excellent choice.

> Homeless need shelter and help. It is inhumane that we let them rot on the streets.

The issue being argued was not that homeless should be able to camp where ever and whenever they want. This issue was that if you don't provide shelter then you cannot kick them out of public spaces.

If that was the law, then it would mean you need to build shelters if you don't want homeless in the streets.

What this new verdict means is:

- You can forcibly remove homeless people who have nowhere else to go

- Thus homelessness can be effectively illegal

- The only realistic solutions now are:

    - Put the homeless in prisons

    - Move them to cities and towns that don't have the resources to remove them.

My experience is that none of these things actually matter anymore. I wish they did, but my experience has been that getting any of these things right in today's environment is in the best case inconsequential to success and in the worse an actual determent.

I've worked at horrifically toxic startup cultures, but it never seemed to hurt them getting funding from outside teams that didn't care in the slightest how healthy the org was or even the product was even remotely good for customers.

I've worked at companies that literally did not know how they were every going to make a profit IPO, then continue to fail to figure this out and changed their strategy to shrug-your-shoulder-and-wait-for-the-return-of-zirp, with no observable consequences on their stock price or investor support for years. Maybe this will eventually catch up to them, but so far they've been quite successful burning a dollar to make $0.70.

I've worked at companies that hired hundreds of completely useless data scientist, filled management with toxic management consultants, driven out all the serious talent they had... with again, no real consequences.

And all of the companies I've known that checked all of your bullet points? Well they've mostly stayed small, saw revenue decline, and eventually start falling apart. All of the best companies I've worked for ended up having to get rid of all the things that made them good in order to appease investors and keep growing.

I wish that we lived in a world where your advice was correct, but I haven't seen any evidence that that is the world we actually inhabit.

Revenue has gone up, but fewer publicly traded companies are making a profit than ever before [0].

I could never understand why so many people just talk about revenue. Revenue without profits is meaningless. There's the old logic of "get enough revenue and then figure out profits and you're highly profitable", but it's very clear that switching the "profit switch" is not so easy in practice.

Investors are still basically waiting for the fed to drop rates, which means that people have abandoned rationally thinking about businesses and are just holding until the free money starts pouring in again.

I honestly don't think the AI bubble is anything like the dotcom bubble. There's something much stranger happening here since the entire market is basically hallucinating and AI is just one manifestation of that.

0. https://finimize.com/content/beware-the-rise-of-unprofitable...

I guess it depends on your definition of "truly gifted" but, working in this space, I've found that there is very little correlation between comp and quality of AI research. There's absolutely some brilliant people working for big names and making serious money, there's also plenty of really talented people working for smaller startups doing incredible work but getting paid less, academics making very little, and even the occasional "hobbyist" making nothing and churning out great work while hiding behind an anime girl avatar.

OpenAI clearly has some talented people, but there's also a bunch of the typical "TC optimization" crowd in there these days. The fact that so many were willing to resign with sama if necessary appears largely because they were more concerned with losing their nice compensation packages than any of their obsession with doing top tier research.

That is fairly well establish to be not true.

"We shall not cease from exploration, and the end of all our exploring will be to arrive where we started and know the place for the first time." - TS Eliot.

You aren't a curious person who studies things for their own sake and finds wonder in exploring ideas. That's fine, this book is clearly not for you so why concern yourself with it? Most of my friends who are obsessed with their field count many intro books as their favorites and frequent re-reads.

> condescending promises of enlightenment.

It sounds like you're more upset that other people enjoy and therefore recommend this book. You're the one asking for proof that it's worth your time. It's clearly not. People who are curious and like to explore ideas in computing recommend this book frequently to other like minded people.

If you don't like sushi, why question somebody's recommendation for their favorite omakase place?

You don't have to read every word of a book to understand if it's interesting to you. I purchased a bunch of technical books the other day that I had never heard of based on opening them up, reading a bit of the intro and flipping through the examples.

Relatively few of my favorite books have come through recommendations compared to those that I have come across through serendipitous discovery.

For anyone who has no time to browse books, then most of the best books in existence would be of little interest to that person.

When I was younger I used to passionately defend those things I've seen as beautiful, but after years experience talking with people passionate about their fields and learning and those who never will be: If you lack the innate curiosity to explore those things others have declared marvelous, then this book will offer you no value.

Every time I crack this book open I get excited and I've read it multiple time and one most of the exercises. I can think of few other books that really expose the beauty and simultaneously strong engineering foundations of software.

You have "tons of experience programming" and sound like you've already decided you know what needs to be known (otherwise why even ask rather than just read it free online), I doubt this will offer you anything you haven't already seen before.

In the visual space, if we're honest, most of the work being done on Stable Diffusion is with the aim of creating "intimate" art. Might not be "companionship" in the family sense, but absolutely a good chunk of the Civit.ai world is, more-or-less, concerned with the mass creation of "waifus".

I don't think that we're going to have many startups succeed in this space, because I suspect the main driver is to have conversations/generate images that people don't entirely want there to be a record off/watched by a private company. But I actually suspect that the number of people using AI for companionship of some form or another is much higher than it appears.

I find this list hard to take too seriously. Despite the city's incredible university culture it's always been fundamentally anti-startup from it's very nature.

Boston is a city were credentials matter far more than ideas, and even then there's a strong, strong culture of "you aren't special, nobody is". People are skeptical of new ideas, compensation for tech has always been abysmal, and people tend to be surprisingly risk adverse.

I've lived there a few times, and there are parts of the city I love, but nobody seriously interested in anything disruptive is going to be hanging in that area too long after graduation.

Each major PC build I've had I've tried to go Linux desktop and ultimately failed because I just got tired of switching to windows for gaming along with periodic cases of essential devices/apps just not working right.

I tried a final time a few years back. I even made sure I had a separate drive just for the Windows partition I felt was inevitable... turns out it never happened.

Gaming on desktop linux is phenomenal now, and I haven't run into any issues with some device or essential app not working. There is no personal task that I ever have to use an old macbook for and it never made sense to add that Windows partition. Occasionally I have to get more fiddly with driver settings to get things to work just right, but not enough for it to be a major annoyance.

Overall if you're tired of your personal OS spying on you increasingly more often and slowly turning into an advertising machine, the linux desktop is ready and quite enjoyable!

If you can't handle "regret" in these cases, then you probably shouldn't be in a position where you're deriving the vast majority of your income/weatlh from investments (which is fundamentally what a CEO does).

It's astounding how many ICs can't wrap their heads around the concept that holding onto your RSUs make absolutely no financial sense. With rare exceptions, this doesn't make sense for anyone. And yet, fear for "regret" keeps people holding.

But it's not shocking that even in tech many ICs are not good at reasoning financially. But if you want to be a co-founder, and hold a lot of your wealth in investments it's essentially that you learn to reason, plan and accept outcomes accordingly. Otherwise you're more-or-less a professional gambler.

I have kids so I'm presuming I'm allowed to have an opinion here.

This is ignoring the fact that babies are not just learning labels, they're learning the whole of language, motion planning, sensory processing, etc.

Once they have the basics down concept acquisition time shrinks rapidly and kids can easily learn their new favorite animal in as little as a single example.

Compare this to LLMs which can one-shot certain tasks, but only if they have essentially already memorized enough information to know about that task. It gives the illusion that these models are learning like children do, when in reality they are not even entirely capable of learning novel concepts.

Beyond just learning a new animal, humans are able to learn entirely new systems of reasoning in surprisingly few examples (though it does take quite a bit of time to process them). How many homework questions did your entire calc 1 class have? I'm guessing less than 100 and (hopefully) you successfully learned differential calculus.