What up nerds, I'm Jerod and this is Changelog News for the week of July 3rd, 2023. Did you catch our conversation with Kelsey Hightower on Changelog & Friends over the weekend? One of my favorite parts was when we were talking about the beauty of a 1.0 and the bloat that occurs when you just keep adding features to software.

[Clip from Changelog & Friends #6](https://changelog.com/friends/6)

We have so much fun on that show. Give it a listen if you haven't yet. And if you dig it, tell your friends about Changelog & Friends. Ok let's get into the news.

Lukas Mathis (author of Designed For Use, one of my favorite design-focused writers) takes up one of the problems with using streaks as a motivator: > losing a streak can be so demoralizing that it can be difficult to start from scratch, and get going again.

I'm in a glass case of emotion!

I've certainly experienced that myself. The worst part: the longer the streak, the harder the fall... enter the concept of "streak redemption" > if your app has a streak feature, provide some way to recover from a streak loss after it has happened. > > This could be a harder task that makes up for a lost day, or maybe a lost day is redeemed if the user manages to continue the streak for a certain amount of days, or perhaps it's something else. I would love the opportunity to double my exercise ring for a day to reclaim yesterday's failure! Let's start working this concept into every piece of software that tracks streaks.

Maybe you've heard [Craig Kerstiens tell us](https://changelog.com/podcast/523#transcript-175) about [JSON in Postgres](https://changelog.com/podcast/417#transcript-36) and how the NoSQL movement came along and had something Postgres didn't have until they added proper JSON support and then Postgres was pretty much back on par with the newer databases.

[Anything you can do I can do better](https://www.youtube.com/watch?v=WO23WBji_Z0)

Looks like it's time to swap out the subject, rinse, and repeat. Here's Jonathan Katz: > Vectors are the new JSON. > > That in itself is an interesting statement, given vectors are a well-studied mathematical structure, and JSON is a data interchange format. And yet in the world of data storage and retrieval, both of these data representations have become the lingua franca of their domains and are either essential, or soon-to-be-essential, ingredients in modern application development. And if current trends continue (I think they will), vectors will be as crucial as JSON is for building applications. There are a bunch of new vector databases on the market, but Postgres junkies should have no fear because [pgvector](https://github.com/pgvector/pgvector) is here... > Similar to the PostgreSQL 9.2 days of JSON, we’re in the earlier stages of how we store vector data in PostgreSQL – and while a lot of what we see in both PostgreSQL and pgvector is very good, it’s about to get a whole lot better.

If finding errors in your backend project is like searching for a needle in a code haystack, Sentry can be your metal detector. Not so sure about that? Join this FREE 20-minute session where two Sentry experts walk you through setting it up for your backend project in minutes. Plus, they're giving a sneak peak of some upcoming error monitoring features. [Sign up here](https://sentry.io/resources/backend-error-monitoring-101/?utm_medium=paid-community&utm_source=changelog&utm_campaign=newsletter&utm_content=newsletter) to learn how you can reduce resolution time down from days to _minutes_ in your backend project. What do you have to lose?! Thanks to our friends at Sentry for sponsoring Changelog News.

Andy Jiang from Deno argues that CommonJS is hurting JavaScript. "JavaScript, the undisputed king of web development, is being sabotaged — not by a rival language or a revolutionary new technology, but by its own baggage from the past. This insidious saboteur is none other than CommonJS, the antique module system that we’ve tolerated for far too long." He goes on to tell the history of CommonJS and how it was born before ES Modules were a thing but now he thinks it needs to go away.

Is it (already) time to scratch "Prompt Engineer" off your resumé? Swyx thinks so: > We are observing a once in a generation “shift right” of applied AI, fueled by the emergent capabilities and open source/API availability of Foundation Models... > I take this seriously and literally. I think it is a full time job. I think software engineering will spawn a new subdiscipline, specializing in applications of AI and wielding the emerging stack effectively, just as “site reliability engineer”, “devops engineer”, “data engineer” and “analytics engineer” emerged. Check Swyx's full article to see why he thinks this is the rise of the "AI Engineer" and why "This will likely be the highest-demand engineering job of the decade"

That's the news for now! Of course, the companion email newsletter features many more stories and links, including: advanced macOS commands, GPT-Migrate, how NASA writes space-proof code, the deep history of who killed Google Reader, and much more. If you're not subscribed to the newsletter pop in your email at changelog.com/news. No spam, no tracking, just the freshens. We have a great interview episode of [The Changelog](https://changelog.fm) for you coming out on Wednesday. Adam and I are talkin' [Efficient Linux at the Command Line](https://www.oreilly.com/library/view/efficient-linux-at/9781098113391/) with Daniel J. Barrett. I've been using Linux for 20+ years and I still learned a lot. Hopefully you will too! Have a great week, **forward this to a friend who might dig it**, and I'll talk to you again real soon.