An open platform for LLMs, speed matters, imaginary problems, Val Town & how to finish your projects (News)

What up, nerds? I'm Jerod and this is Changelog News for the week of Monday, June 19th, 2023. Did you catch our "Mat Depends" episode of Changelog & Friends over the weekend?

Early reviews are in and glowing!. Here, I'll read you a few. "Absolutely delightful! The mix of informative content and humor is just perfect." "This is the kind of content that keeps the developer community buzzing – kudos to the team!" "Perfect balance of humor, insightful discussions, and informative tips – an absolute winner." And thats just some of the responses ChatGPT hallucinated when I asked it for a bunch of glowing reviews. Ok, let's get into the news.

Just when you think ClosedAI & big tech might dominate our inevitably AI-infused future... open source enthusiasts all around the world rally around each other's work and launch [project](https://gpt4all.io/index.html) after [project](https://falconllm.tii.ae) after [project](https://github.com/bentoml/OpenLLM) offering viable (and sometimes superior) alternatives. Enter OpenLLM: a platform for operating LLMs in production. With OpenLLM, you can run inference with any open-source large-language models, deploy to the cloud or on-premises, and build powerful AI apps. I've said it before and I'll keep saying it 'til it's not true anymore: always bet on open source.

Back in 2015, James Somers wrote a post that is still relevant today. It's called Speed matters: Why working quickly is more important than it seems. In it he says "The obvious benefit to working quickly is that you’ll finish more stuff per unit time. But there’s more to it than that. If you work quickly, the cost of doing something new will seem lower in your mind. So you’ll be inclined to do more." That's an interesting way to think about it. He likens it to email exchanges: "I’ve noticed that if I respond to people’s emails quickly, they send me more emails. The sender learns to expect a response, and that expectation spurs them to write. That is, _speed_ itself draws emails out of them, because the projected cost of the exchange in their mind is low. They know they’ll get something for their effort. It’ll happen so fast they can already taste it." He generalizes this phenomenon as "systems which eat items quickly are fed more items. Slow systems starve." and gives a couple more examples, if you're not convinced.

Here's another solid piece of writing that had me nodding along in agreement: Imaginary problems are the root of bad software. Author George states: "There are many factors which can be a catalyst for bad software: from the tools being used, to team communication, to the personal stake developers have in its success, to the testing methodology. I propose that there is one problem chief among them, an impetus for bad software from which almost all others take root: imaginary problems." This is why I say YAGNI ad nauseam and fight tooth and nail not to add new features until their value is already demonstrable.

Sponsored News time, y'all. Two weeks back I told you about HasuraCon and a couple listeners liked the ad so much they tweeted about it. So, following the DRY principle I'm going to call the same function and just pass it some new values. Here we go. Our friends at Hasura would like to invite you to HasuraCon on June 20th through the 22nd! Three days to learn, share, celebrate, and geek out on the future of Hasura and data APIs. Data APIs are reshaping the world of data delivery, helping enterprises do more with their data by serving it where it’s needed and when it’s needed in a fast, secure, flexible way. Hasura and their customers are at the forefront, driving this shift. Costco, Verizon, Atlassian, General Mills, over 40 of the fortune 100 companies use Hasura. HasuraCon features a world-class speaker lineup, hands-on workshops, product deep-dives, and more. Whether you’re a seasoned Hasura pro or just starting, there will be something for everyone. And best of all, it's 100% FREE and easily accessible online. Get all the details and register today at http://hasura.io/hasuracon Once again that's http://hasura.io/hasuracon

Val Town is a new social website to write and run code. The tagline for this brand new site from [Steve Krouse](https://stevekrouse.com) (bankrolled by [Daniel Levine](https://www.accel.com/people/daniel-levine)) hits the bullseye: If GitHub Gists could run And AWS Lambda were fun Gists are cool. But they'd be cooler if you could execute them. Lambda is cool. But it be cooler if it were actually fun to use. Enter Val Town, where you can write and share "Vals" (Vals are small JavaScript or TypeScript snippets of code, written in the browser and run in their servers.) It's still early days for the service, but it's a cool idea with a cool domain. Check it out at val.town

Aaron Francis for GitHub's ReadME Project: > Sarting a new project is a rush. The possibilities are infinite. There's no legacy code dragging you down; we're only making good decisions this time! The beginning of any project is always characterized by blissful productivity.... > > Sooner or later, the blissful productivity gives way to something that feels much more like... work. More like a grind. But it's probably just this project, right? You've lost interest. The passion is gone. It's not as fun as you thought it would be. All that's left is the "boring" stuff. > > You have a new idea, though, and you're sure that you'll see this one through! > > And so the cycle continues, over and over again, until you're left with a graveyard of unfinished projects, wondering how anyone ever finishes anything. What does everyone else know that you don't? Starting something is easy. Anybody can do that. Finishing is hard. It's what sets people apart. In this guide, Aaron walks you through the work of finishing and helps you take your project from "mostly done" to "actually done"

That's the news for now! Wednesday's interview is with frequent guest [Adam Jacob](https://changelog.com/person/adamhjk) who's been toiling in stealth mode on what he calls the 2nd wave of DevOps for _years_ and he's finally ready to talk about it. On Friday's talk show we're joined once again by our old friend [Brett Cannon](https://snarky.ca/author/brett/) from the Python Steering Council who would really like to help me with my pip install anxiety. Have a great week, forward this (or [copy/paste](https://changelog.com/news/49/email)) to your friends who might dig it, and I'll talk to you again real soon.