The Changelog: Software Development, Open Source - The AI 2027 scenario (News)

Episode Date: April 7, 2025

Daniel Kokotajlo and the AI Futures Project lays out a potential scenario of superhuman AI's impact, Liam ERD generates beautiful, interactive ER diagrams from your database, Mozilla takes on Gmail with "Thundermail", algernon explains why grepping remains terrible & Vitor M. de Sousa Pereira rans on the insanity of being a software engineer.

What up nerds? I'm Jared and this is changelog news for the week of Monday, April 7th, 2025. Now that Severance Season 2 is all wrapped up, I need something else to watch of similar caliber and intrigue. Otherwise, I'll fall back into my New Girl and American Pickers reruns habit. Hit me with your recommends. Okay, let's get into this week's news.

The AI 2027 Scenario. The AI Futures Project is a new nonprofit organization led by Daniel Kokotajlo, a former researcher in OpenAI's governance division with a goal of forecasting the future of AI. In their first publication, AI 2027, the group predicts that the impact of superhuman AI over the next decade will be enormous, exceeding that of the industrial revolution. They lay this potential future out in a scenario that quote, represents our best guess about what that might look like.

It's informed by trend extrapolations, war games, expert feedback, experience at open AI, and previous forecasting successes." What's cool about this work, unlike so many that you can find online, is its incredible detail. The story goes quarter by quarter through the next 2.5 years, weaving one potential narrative

of how things might proceed. There are two endings though, a slowdown and a race ending. The overall goal is not prescriptive, it's predictive accuracy. And they're so serious about it that they're offering cash incentives for alternative scenarios that prove out to be more accurate. You can find a link to the cash incentives in the newsletter. Liam ERD. Quote, Liam ERD generates beautiful interactive ER diagrams that's entity relationship diagrams from your database.

Whether you're working on public or private repositories, Liam ERD helps you visualize complex schemas with ease. End quote. Pretty cool project. You can get started using Liam with zero config. Just provide your schema and you're good to go. It currently has direct support for schemas from Postgres, Ruby on Rails, Prisma, and TBLS.

Other databases are in the works and many have workarounds for the impatient. Mozilla Thundermail takes on Gmail. Mozilla is turning its Thunderbird open source email client into a full communications platform with the launch of Thundermail and Thunderbird Pro. The expansion of Mozilla's email services aims to compete with rival ecosystems like Gmail and Microsoft 365, which are more rich in features, except Mozilla's offering stands out for its open source values of privacy,

freedom, transparency, and user respect." End quote. Thunderbird has been losing users over the last five years, but they're hoping the launch of Thunderbird Pro reverses that course. Pro consists of a few different subservices. Quote, Thunderbird appointment, a new scheduling tool for sharing calendar links.

Thunderbird Send, a rebuild of the discontinued Firefox Send, and Thunderbird Assist, a new AI-powered writing tool enabled via a partnership with Flower AI that is intended to do the processing locally to eliminate privacy concerns. End quote. Once the full hosting service is ready, users will get to pick between

thundermail.com and tb.pro domains for their email. Both of those sound like duds to me. How about Thundermail for domains? That'd be pretty cool. It's now time for sponsored news. New Depot registry makes builds even faster Faster Everyone knows you can increase the speed of your builds with a one-line code change thanks to Depot. But did you know you can also leverage their new registry to make your builds even faster?

Depot has revamped their ephemeral Docker registry and promoted it to a full product called Depot Registry. It's engineered to serve images from the most optimal location, no matter where in the world the client is located. Here's a note from co-founder and CEO, Kyle Galbraith, quote, "'Depot Registry is a new, faster, and more feature-rich registry

that we are excited to keep iterating on. It's available to all users starting today,' end quote. The Depot Registry automatically serves images from the nearest location worldwide, ensuring lightning fast performance for every user. By leveraging tigress for intelligent content routing and replication, depot seamlessly replicates each layer to 13 global storage regions, acting as both a CDN and object storage provider with an S3 compatible interface. If you are new to Depot, you can get started with Depot Registry

by signing up for a seven day trial. No credit card required. Find the link in your chapter data and the newsletter. Grepping logs remains terrible. Hacker Algernon first stated his public opinion that grepping logs is terrible a decade ago. The good news is he's back and the bad news is his opinion has not changed. After a quick comparison of running alight queries against 4.4GB of JSON logs on his

modern M2 SSD desktop and loading the same data into Victoria logs on an old Mac mini, he finds, quote, what we saw here is that a computer with much more RAM and computing power and faster storage got absolutely obliterated by an overloaded Mac mini that is 10 years its senior, and not only in speed, but resource use too. I haven't shown it, but working with raw text

used a lot more CPU time, disk IO, and memory. Meanwhile, the Mac Mini barely blinked. And not because it was all cached in memory. These are all cold queries. Cached queries are much, much faster. But how can it be that such an underpowered device runs circles around a powerful desktop? Purpose built software, dear reader. Purpose built software, dear reader. Purpose built software."

End quote. The moral of the story, databases are good. Quote, don't grab your logs, shove them in a log ready database and query that. Not only will it require considerably less disk space, it will require less computing power, less RAM, and as a result, less time to perform the queries. And if that's not enough, you can give the logs structure.

You no longer need to remember the order of fields in HTTPD's common log format. You can query against a named field instead. You can run aggregations, computations, build stats, correlate, join, and do all kinds of other database-y stuff. Not only does it perform better, you can do more and do so more conveniently. The Insanity of Being a Software Engineer Vitor M. D'Souza-Periera pens a solid rant

on the never-ending list of technologies that software engineers must master and skills that we must acquire. Quote, A recruiter reached out to me a couple days ago about an engineering position for a secret company. They decided that they required senior level skills in rails, hot wire, and incredibly native mobile development. Why not add kernel and compiler development in there as well for good measure?

When a house is being built, tons of people are involved. Architects, civil engineers, plumbers, electricians, bricklayers, interior designers, roofers, surveyors, pavers, you name it. You don't expect a single person or even a whole single company to be able to do all of those." End quote. VTOR's post focuses on web development,

a subsection of software engineering where this pain is particularly acute. But I've long said that if you don't want to be continually learning new things to sustain your career trajectory, you don't belong in the technology industry. Still, it can help to air our grievances once in a while with a solid rant like this one. That's the news for now, but also scan the companion newsletter for even more links worth clicking on such as… Self Hosting like it's 2025 A Typeface for creating sparklines in text

without code And the blissful zen of a good side project Get in on that newsletter at changelog.com news. Have a great week, like, subscribe, and review if you dig the show.