alexmuro

I generally am very satisfied with fastmail but I do have a hard time getting it to learn what is spam, maybe I'm missing something.

AI has definitely shifted my work in a positive way, but I am still playing the same game.

I run a small lab that does large data analytics and web products for a couple large clients. I have 5 developers who I manage directly, I write a lot of code myself and I interact directly with my clients. I have been a web developer for long enough to have written code in coldfusion, php, asp, asp.net, rails, node and javascript through microsoft frontpage exports, to jquery,to backbone, angular and react and in a lot of different frameworks. I feel this breadth of watching the internet develop in stages has given me a decent if imperfect understanding of many of the tradeoffs that can be made in developing for the web.

My work lately is on an analytics / cms / data management / gis platform that is used by a couple of our clients and we've been developing for a couple of years before any ai was used on it all. Its a react front end built on react-router-7 that can be SPA or SSR and a node api server.

I had tried AI coding a couple times over the past few years both for small toy projects and on my work and it felt to me less productive than writing code by hand until this January when I tried Claude Code with Opus 4.5. Since then I have written very few features by hand although I am often actively writing parts of them, or debugging by hand.

I am maybe in a slightly unique place in that part of my job is coming up with tasks for other developers and making sure their code integrates back, I've been doing this for 10 years plus, and personally my sucess rate with getting someone to write a new feature that will get use is maybe a bit over 50%, that is maybe generous? Figuring out what to do next in a project that will create value for users is the hard part of my job whether I am delegating to developers or to an AI and that hasn't changed.

That being said I can move through things significantly faster and more consistently using AI, and get them out to clients for testing to see if they are going to work. Its also been great for tasks which I know my developers will groan if I assign to them. In the last couple months I've been able to

- create a new version of our server that is free from years of cruft of the monorepo api we use across all our projects. - implement sqlite compatablity for the server (in addition to original postgres support) - Implement local first sync from scratch for the project - Test out a large number of optimization strategies, not all of which worked out but which would have taken me so much longer and been so much more onerous the cost benefit ratio of engaging them would have been not worth it - Tons of small features I would have assigned to someone else but are now less effort to just have the AI implement.

  I think the biggest plus though is the amount of documentation that has accrued in our repo since using starting to use these tools. I find AI is pretty great at summarizing different sections of the code and with a little bit of conversation I can get it more or less exactly how I want it. This has been hugely useful to me on a number of occasions and something I would have always liked to have been doing but on a small team that is always under pressure to create results for our clients its something that didn't cross the immediate threshold of the cost benefit ratio. 

In my own use of AI, I keep the bottleneck at my own understanding of the code, its important to me that I maintain a thorough understanding of the codebase. I couple possibly go faster by giving it a longer leash, but that trade off doesn't seem wise to me at this point, first because I'm already moving so much faster than I was very recently and secondly because it doesn't seem very far from the next bottleneck which is deciding what is the next useful thing to implement. For the most part, I find the AI has me moving in the right direction almost all the time but I think this is partly for me because I am already practiced in communicating the programmers what to implement next and I have a deep understanding of the code base, but also because I spend more than half of the time using AI adding context, plans and documentation to the repo.

I have encouraged my team to use these tools but I am not forcing it down anyone's throat, although its interesting to give people tasks that I am confident I could finish much quicker and much more to my personal taste than assigning it. The reactions from my team are pretty mixed, one of the strongest contributors doesn't find a lot of gains from it. One has found similar productivity gains to myself, others are very against it and hate it.

I think one of the things it will change for me is, I can no longer just create the stories for everyone, learning how to choose on what to work on is going to be the most important skill in my opinion so over the next couple months I am going to be shifting so everyone on my team has direct client interactions and I am going to try to shift away from writing stories to having meetings where I help them decide on their own what to work on. Still part of the reason that I can afford to do this is because I can now get as much or more work done than I was able to with my whole team at this time last year.

That's a big difference in one way, and I am optimistic that the platform I am working on will be a lot better and able to compete with large legacy platforms that it wouldn't have been able to compete with in the past, but still it just tightens the loop of trying new things and getting feedback and the hardest part of the business is still communication with clients and building relationships that create value.

I read the Wikipedia on xslt, and as a long time web developer i do not understand at all how this would be useful. Plenty of people here are saying if this tech had taken hold we'd have a better world. Is there a clear example somewhere of why and how?

Thank you for the link to the git repo, this looks great. Thank you for your work. MapLibre is a library I use all the time and while MVT isn't something I have any complaints about this will still be a big upgrade.

This is interesting, is there a reference implementation that exists somewhere? Will there be a fork of tippecanoe that can encode these files or something different?

I think closed vs open problems are not simply different in magnitude of difficulty but qualitatively different. When I'm programming most of the interesting things I work on don't have a clear correct answer or even a way of telling why a particular set of choices don't get traction.

I guess it's possible that just being "smarter" might in some cases get a better solution from a seeies of text prompts but that seems too vague an argument to hold much water for me.

I don't feel like we are in the waning days of the craft at all. Most of the craft is creating an understanding between people and software and most human programmers are still bad at it. AI might replace some programmers but none who program as a craft.

As someone who creates data and analysis which get used in setting policy I do find a lot of EA spreadsheet analysis of measured "good" to be very niave to the nature of measurement and classification.

That being said, I think this peice is a bit of an overreaction and there seem to be many earnest actors in the EA community really thinking about how they can do good in the world. SBF is very unfortunate for EA, but to jump from him example to saying all EA practitioners care exclusively about the ends over the means is a bit of a leap, imo.

This is one of the best things I've read on the internet in a minute. It's a modern Koan.

Just use your freewill.

Delightful.

It's crazy how little has changed and how much had changed. I remember what a revelation "the unreasonable effectiveness of RNNs" was when I was read it and it feels like we live in a different world.

Firstly, I just want to say thanks for the reply, but more so thanks for your work, its moving opensource mapping forward.

In my work we are looking at switching from mbtiles hosted with tilserver-gl(https://github.com/maptiler/tileserver-gl) to pmtiles to remove a server process. But we were self hosting already and we are already using maplibre-gl 2.

I can see why the implementation in the blog post would be better for high traffic deployments (ours isn't). It also points out to me I don't understand how a CDN would handle range request for hosting the pmtiles file directly, it probably doesn't?

As far as the mapbox stuff,in my mind, pmtiles is a direct competitor (successor) to the mbtiles format, which was a revolution in comparison to everything that came before it. A successor I welcome because it makes it even easier for me as a developer to self host and not be dependent on a SaaS to run my maps.

The modern opensource map stack wouldn't exist without mapbox and I'm personally grateful to them for that. Most people who use pmtiles will use mapbox's opensource style spec to style them, and descendants of their open source code to render them. But as a developer now its an obvious choice to not use their services after years of using them.

However I'm not doing high traffic stuff and they never made much money off me anyway.

The thing this article doesn't say is that maplibre-gl v2 supports directly querying pmtiles with http range requests so you don't even need lambda or cloud flare workers to make x/y/z routes in front of the file. So instead of 50c, this is essentially free.

If you are going to to set up that infrastructure you could just use an mbtiles file which has been around for years.

The interesting thing to me is that this stuff is all built on the open source technology of mapbox, and it seems like a real threat to large parts of their business model. Interested to see how it plays out.

This is a great article but it left out the side of this that I was really interested in which is the compute performance of these different options.

I'll take a fast deep clone for my case over one that handles all cases.

This will directly lead to me no longer using signal. What are other people switching to for their default sms client on android?

I used to deploy SPAs like this and I consistently had problems with cloudfront's cache clearing. Projects would not update for random amounts of time and at different times for different clients. I thought I was pretty good a trouble shooting stuff like this but over months of effort I could never get it to work update when I invalidated the cloudfront cache no despite trying many different approaches.

In the end I have switched to hosting on Netlify, which is easier to set up, not that I particularly care about that, but when I deploy my live site updates immediately.

I think the perspective of this article is a very healthy and productive one and I found it particularly useful.

Assessing if attending on the shoulders of the new Giants who stand up every few years is a difficult problem that I'm interested in and I appreciate the amount of context given here considering different use cases.

The best thing that you can do to give a project legs is to get other people involved. Even if you don't have someone who can work on coding with you, just sharing it and enrolling other people in your vision for a project is the number one way to give you the motivation to continue working on it.

Personally the biggest advantage .NET core has for my situation is that it creates a pool of programming talent and mindshare that I'd now be open to hiring and working with on projects that can now be made with open source tools in a sane environment.

I don't think it would be my go to choice for projects currently, but its an option I would happily consider and I give Microsoft a lot of credit for moving in this direction.

That being said so far the reasons I've had to look at it were modernizing older .NET projects and last I looked there were so many incompatibilities be between .NET core and .NET standard that the projects would have been almost complete re-writes, somewhat of nullifying its advantages against other choices, but that was a while ago.

tldr; I've invested a lot of time becoming very good at rails and I don't want to learn a new appraoch to solving a problem I am already very good at solving.

The quality of an end product doesn't have as much to do with tools as people would like to believe.

Personally I recommend Stanford CSI 231n http://cs231n.stanford.edu/

Its specifically geared towards visual recognition, but it starts with the basics of machine learning and moves on to feed forward nets and covnets and covers RNNs and attention towards the end.

The assignments are a great set of jupyter notebooks that really get your hands on the material and you can find a number of peoples complete assignments on github just by searching.

The lectures are available online as well https://www.youtube.com/playlist?list=PL3FW7Lu3i5JvHM8ljYj-z...

I've done hinton's and Ngs courses and as someone who already has a non-ai development background I found this to be the best introduction. Its really an extension of Andrej Karpathy's Neural Nets for Hackers (http://karpathy.github.io/neuralnets/)

I know and the piece is spot on, not to mention the fact that with PostGIS its the best database for spatial anything, much better then expensive proprietary solutions. Thanks postgres!

just want to say thanks to all involved in this interview. This is exactly the kind of discussion that imho can change minds and move this issue forward. I have so much respect for these women.

I wrote a pretty big app in falcor, using react-redux-falcor for the client. Overall its was a good experience.

I haven't really gone back to check out graphql thoroughly, although I will say since I started my falcor project the community around graphql has grown a good deal faster than the falcor community.

I haven't worked on my falcor backed app in a couple months but last I worked there were still some rough edges and small bugs with falcor and it hadn't made it to a 0.1.0 release yet and there wasn't a lot of info from netflix about when it would get there.

If I was going to start a new app right now I'd take a longer look at graphql, despite really liking the development experience of falcor because of the community / long term outlook.

I am getting a 503 for all their js files hosted on the cdn.melowdown.com

If you are doing a simple project, there is nothing wrong with just using jquery (or imho d3). If you know that your project is always going to be pretty simple or especially if you are learning.

I have been teaching a number of people javascript at my lab lately and I think its unwise to teach new javascripters using any of the modern tools or transpiling until someone runs into one the problems that those tools solve. Having been a javascript coder consistently from the jquery days until now, I am not pining for the simple old days of yore.

It can be hard to learn everything that goes into a big front end app these days but when used correctly all those tools are there for a reason.

You can also wrap the reactjs web application in in Electron (http://electron.atom.io/), which supports Mac, Windows and Linux.

imho, this is worth showing as a piece of design.

Its an innovative calendar layout and while I don't know if I would use it in time / data selection, in terms of visualizing data inside of a year it has a lot of potential.

here is a link to the google cache.

http://webcache.googleusercontent.com/search?q=cache:d0Df-gF...

Jafar Husain does an excellent job of explaining some of the problems netflix has had with REST APIs and why you might want to take a different approach in his introduction to Falcor video. Falcor is tackling some of the same problems as graphql / relay but making trade offs at different points.

https://netflix.github.io/falcor/starter/why-falcor.html

A REST API may be the best decision for a lot of projects but GraphQL and Falcor are solving interesting problems that come up more and more in single page apps of increasing complexity.

This is timely again, putting out new hardware and integrating every device with windows, even open sourcing its development platform isn't going to bring the developers back from the web.

Personally I don't think it was as obvious in 2007 as it is now, that all apps would be on the web. Whats stranger is that Microsoft still doesn't seem to understand this.