Loading…

Issue 316: A technological 2-for-1

It’s time for a classic Stack Overflow Q&A. Q: What’s better than one interview from the floor of re:Invent? A: FOUR interviews from the floor of re:Invent. Also, this question is now closed for being off-topic. Okay, okay, fine, let’s try to stay on-topic—namely the topic of AI. On the pod this week, we’re bringing you chats with Inception’s Stefano Ermon on the power of diffusion models and Roomie’s Aldo Luevano on building physical and software AI with a purpose and real ROI. We’re also joined by Pathway’s Zuzanna Stamirowska and Victor Szczerba to dive into the world’s first post-transformer frontier model, and Mary Technology’s Rowan McNamee to chat about LLMs in the legal world—we’ll have to ask him if this week’s 4-for-1 podcast deal is so good it should be illegal.

Speaking of The Law, we consult the Law Stack Exchange as to whether social media grifting is grifting at all—plus the answer to your burning question on what happens to rocket ship boosters that don’t burn up. Not everything is rocket science, though. For instance neural networks, especially since we have a visualizer for you this week that’ll demystify those mystifying robot brains. Let’s stay on-topic and continue our demystification—we’ve got the story of one dev’s attempt to find what’s on the other side of Google’s 8.8.8.8 DNS. Maybe we owe all the mysteries around the tech we build to the complexity we’ve been adding to it, which is probably why one of the stories from the web this week is on Wirth's Law of lean software.

Oh no, we’ve gone off-topic again. Well, we tried our best. And don’t worry, we’ve got plenty of on-topic and not-closed questions to round out this week’s off-topic Overflow. Is impersonation the highest form of flattery if you’re impersonating a Window’s user with lower privileges? Can you “just bumping this thread!” and “quick follow-up on this!” your way into faster code review? Should you let AI kill your darlings if your darlings are all trash? All of those wonderfully on-topic answers ready for you in the links below.

From the blog

AI attention span so good it shouldn’t be legal

We have another two-for-one special this week, with two more interviews from the floor of re:Invent.

Code smells for AI agents: Q&A with Eno Reyes of Factory

Quality software still needs high-quality code, AI agents or not.

Generating text with diffusion (and ROI with LLMs)

Two guests for the price of one! This episode has two interviews recorded at AWS re:Invent back in December.

Interesting questions

Can people who donated to someone under false pretenses get their money back?

Old school fraud in a new school, social media grifter world.

What actions can I take as a junior developer to solve long review delays?

Maybe you'll "per my last email" your way into a senior developer role.

What happens to SLS boosters when they are expended in the ocean?

What do the Artemis II test flight and the Super Bowl have in common? You can sports bet on both.

Links from around the web

Stay away from my trash!

To fight AI PR slop, sometimes you have to be like Danny DeVito and say, "I'm the Trash Man!"

Understanding neural networks, visually

To understand AI, you have to understand your own mind.

What's at the other end of 8.8.8.8?

Why was 6 afraid of 7? Because 7 8(.8.8.8) a Google-owned IP address.

Wirth's revenge

This just in: just because software can do it for you doesn't mean it should.


We want to know what you're using to upskill and learn in the age of AI. Take this five minute survey on learning and AI to have your voice heard in our next Stack Overflow Knows Pulse Survey.