I get what the author is saying, but his examples like "a ripgrep fork that ... is faster for my codex ripgrep workloads" don't strike me as particularly ambitious.
Perhaps we should be more ambitious in the types of problems we try to solve, and the types of value we try to provide with the software we build.
Amen to that. I was surprised to see yours seems to be the only comment with this sentiment. This post was kind of depressing to read. Like if AI is so amazing and can one-shot so much astounding stuff, why is anyone wasting their time on pointless "projects" like making ripgrep faster?
Because it can cut their total time by double digit percentages?
Both your comments there seem to latch on to a detail and miss the context totally. Ambition scales with the number of "small" things you can get done too.
Total time of what? Like what is the end goal that people are trying to achieve? All of the examples I can see in the post are similarly about development-internal stuff like tests and performance (or games). What I mean by true ambition would be stuff that is actually aimed at helping people in the world at large, not just helping developers develop.
Ambition doesn't scale with the number of small things you get done. Ambition is the size of the big thing you are trying to do. It's true that improving small things can help you do that big thing, but that's not the same as the ambition itself. Wanting to build a better mousetrap is just a modest technical ambition. The kind of ambition I'm talking about is wanting to eradicate the plague, and then maybe building a better mousetrap as a means to that end.
Indeed, the author has written on multiple previous occasions about how small efficiencies compound. I think it doesn't always work for humans (for psychological reasons: e.g. if we get to start out lunch break a few minutes earlier it won't translate into coming back sooner too), but there's something to it.
Ambition describes the big things you intend to accomplish. By my definition, this is the opposite of solving a lot of small inconveniences. If all you intend to accomplish is to optimize your personal microbenchmarks, then you're primarily navel gazing.
The note taking genre of opensource apps is saturated with many options but most of them are electron based.
I was hoping that someone with enough motivation and money for tokens would come up with a non-electron option that is similar to onenote, with free form note placement and handwritten notes.
Basically I wish there is an opensource app having good feature parity with OneNote. Thats a good benchmark to have.
A few years ago I was wishing the same, but I was wishing it for someone knowledgeable enough.
Today I wish the same, but for someone with enough motivation and tokens instead.
Don't get the wrong idea: I would be glad for a OneNote/Obsidian alternative that's performant. I just lament the days where I'd stare in awe thinking about the levels of wizardry one must get to in order to build something that I admire. Now the first thing that comes to mind is "I wonder how many tokens did this took them to make".
an ambitious version of a note-taking program might be one that comes in pieces that you then put together by explaining to an AI what type of program you want, adding only the features you want and ignoring everything else
its always been a problem for me that ive never been able to find something that works exactly how i want, so i just put up with whatever is close enough, like obsidian for example. its probably because note taking is just a very personal thing so everyone is going to have different needs, and the amount of options out there is kind of proof of that
from the point of view of a developer, an AI would help with the complexity of a program that has every feature under the sun, and for the end user ian Ai would help the complexity of having to put all those modules together and making them work well
obviously each person could just vibe code their own program right now but each one would be starting from scratch. it would be more efficient to start with the same modules that have been tried and tested and then piece them together into something more standardised
i would love if this was a thing for most software actually. photoshop for example has only a dozen features that i ever use and the other 1000 features are wasted on me and just mean a longer startup time and a larger install size, but also i dont want want an AI to make me a custom version from scratch because you would probably spend more time trying to iron out all the weird bugs in it than actually using it
Then I had a form that let you pick which ones you want. It spit out a gm script with all entries commented out except from the ones you wanted to keep.
When a link or button is clicked the server return a js file that contain one function. Then it executes the function.
To borrow your example, when you've picked the functionality you want in your note taking app a version number like 1.0.0.unixtime can be generated and stored so that you can later populate the config form with the chosen configuration and add or remove things.
Or you could do something with a config file to populate the form again.
An LLM can also modify a config file. Before execution it can be validated, as it is so simple it can show which mandatory entries are missing, which ones are spelled incorrectly.
If you provide the LLM with the html form or the config file (+docs) it shouldn't have much difficulty generating exactly what you wanted.
As the application has now been turned into ravioli it also addresses the ancient puzzle of separating feature updates from security updates.
You can still introduce a new gui that confuses the hell out of users. Then if no one checks the box you can take your learnings from it without 1000 people having to tell you in 500 different places.
you don't like the ads, gambling apps, and similar? /s
What might help is non-software-first companies valuing developers and engineers more, enough that we could earn a reasonable living which building actually cool things. Not sure how the economics works out on the other side. Can these companies make enough money to support this vision?
One major obstacle is developers have either encountered or created entire ecosystems of problems they can't solve that get inherited whenever and wherever software is introduced. Proposing to add evergreen issues around security, data entry and retention, connectivity, vendor lock-in, and update cycles is frequently a shitty trade for whatever notional (often fictional) productivity improvement is being proposed by injecting software into a workflow.
My building recently added ButterflyMX and boy does no software sound so much better. This is tech company that tracks people as the move across real world properties and sells your data to the ad industry.
I mean they're an excellent example, but this is just as true for the overwhelming majority of B2B productivity suites I've encountered. Hell, digital cameras didn't improve photography, they just killed three industries and saddled just about everyone with data retention and migration problems and that's without even getting into the social effects of ubiquitous cameras combined with social media or the surveillance state they've made possible.
More accessible? Barely. Kodak disposables and Polaroid cameras were so common they were frequently used as toys by children. Also you can't claim zero marginal cost. If anything continuous data retention headaches make it more expensive (in terms of time) to take and keep photographs than any point in prior history. The costs just moved from the camera to the photo album.
Continuous retention is a huge issue that needs to be solved, but it is not insoluble. I would not be surprised if most kids are losing their childhood photos from services being shut down, tinkering with quality, or altering terms.
However the rest of the points are wrong as the photo album step was there originally. Film acquisition time, album acquisition time, labeling time, sorting time, sharing time, shelf space creation and cleaning time have all improved by orders of magnitude and all of that impacted accessibility.
Film acquisition time? You tossed it in the shopping cart while walking through the grocery store along with batteries and other random household essentials. In extremis you'd grab a few rolls off the display in the closest gas station. People literally spend more time a year getting off their phone lock screen to take a photo than you'd spend in the act of buying film. You're talking about a task with seconds-per-year levels of overhead for normal households. Professional photographers ordered theirs in bulk shipments so again the unit cost in time was functionally nil. Same with albums. Labeling and sorting were optional and at most happened one time (1) per pack of proofs and this was typically a group recreational activity for families. Dust gathers at a set rate regardless of what is occupying a given space, so it's not obvious how having a photo album increases cleaning time.
As to the claim that retention isn't insoluble, that's at best a baseless statement of faith not terribly dissimilar to someone praying a rosary. With other religious pieties there's at least some evidence (someone built a cathedral, there's a book, etc.) to support the faith. There is no evidence to date to support the idea that basic file formats are a settled matter on archival timeframes, much less retention services or storage hardware.
I would argue the camera becoming an app on a phone has done this more than digital cameras. You don't need a separate device, you always have it with you.
A big benefit here is public transparency and citizen video evidence. Like when ICE murdered Alex and Renee
I'm all about ending user tracking, targeted advertising, and gambling. We'll need to find new things to work on that do not use big data to take advantage of our fellow humans.
Gell-Mann amnesia doesn't really apply to the world expert on a topic discussing a paper that builds on his own work. It might apply to a world expert on a topic discussing a topic on which he is not an expert.
You're focusing too much on semantic nitpicking and less on the part where people "wield it as a tool to fuck over artists."
Nobody cares if hobbyists make a "Star Wars but make it Monet" art piece with AI and put it on their wall. It's the part where bosses fire John Smith and then ask the LLM, "please write 100 articles in the style of John Smith." It's the part where an actor is hired once, their face scanned, and then their likeness is used (unpaid) in a hundred films.
It might have been the Exponent podcast years ago that was talking about how a set of contracts, practices, and (maybe) laws should be drawn up to cover that exact face-scan-the-actor use case, but enable royalties to go back to the actor. So then Shaq could appear in so many more commercials (lol)
reply