Surprise (/s): Software engineering is changing again.

Since the dawn of software as we know it, we've been able to react to these changes and adapt to them. First it was the Web 2.0 and WYSIWYG, then it was the huge cloud push and SaaS, then we had a little Web3 sidequest, and now in the end game, it's AI hype and cinematic launches.

We've had our fair share of weird crazy sh*t that's made one go what the fuck. (for me, it's the YouTube adblocker war) -- that's consistent. However, that same consistency is where it all starts to get weird.

Hiring

Firms are now using AI ATS to screen AI chat written resumes, to interview AI assisted candidates, who, when hired will write code in—you know the rest. And I hate when people blame the candidates because the tools are too good. They're then doomed to using AI to rewrite their resumes entirely. And if they don't, they'll end up back at square one. With no job. It's a Linked List Cycle, literally.

Like look. If you've applied anywhere recentely, you know Ashby:

Need I say more?

And since GPT is no longer bad at this like I said with the tools being too good, it's now become an arms race to the bottom. Since the reality is that as time goes on, companies will only have more and more applications for shrinking single digit positions; claiming to want an expert in every facet of a given job, not realizing that early-career engineers are always learning new things. It's a game of "who can write the best AI-generated resume".

A case study: Porn resume lands interview(s) because ATS is so good right?

https://www.reddit.com/r/recruitinghell/comments/qhg5jo/this_resume_got_me_an_interview/

This is why I believe companies are going to start going more and more by pedigree. The tech bros have successfully become the finance bros. When it's near impossible to filter down candidates, they'll just learn at the good old reliable pedigree. School name, past internships, all the stuff that makes it harder for people like me to even enter the fray.

At least if someone went to MIT or Stanford, you can at least trust they have some skills since they had to prove themselves to get in. All the other ways in are becoming unreliable. Cold reach out? Someone already crafted a targeted email to the CEO. Made an impressive OSS project? Someone sloppified it and used it on their own CV.

Fake open source contributions

This has always been a problem. HOWEVER, now people are using AI to complete the E2E action of a PR. It no longer takes 5-50 minutes out of a fake contributor's day to open an issue and PR. Now they can clone it locally, ask AI for some random improvement, and submit it acting as if the code was bread becoming wine.

The good news? People are noticing. One of my favorite videos from Theo Browne, a founder of T3Chat, who made this video below, is I think; given his position in the full-stack community, a great case of someone using their platform to shed awareness and call to action change.

I've been made aware of PRs like this one to Node.js that are clearly AI-generated - the code, the commit messages, even the PR descriptions. The maintainers caught it and called it out, but how many slip through?

11M seed round with no users?

Dedalus Labs raised a $11M seed to quote "redefine how developers build AI agents".

Now normally, I'd be all over this. But upon looking deeper, I noticed some things.

1. They're not any different from langchain, Vercel AI SDK, or really..any other AI SDK.
   • Their founder said that they don't just "route models; we support vendor agnostic model handoffs" when asked about their difference from langchain. But that's part of the problem. Langchain DOES do that! Vercel AI SDK DOES do that!
2. The founder said they support python and go as a selling point...but sure, more python SDKs for building agents are exactly what the ecosystem needs! /s
3. Their most basic example is not even complete!

Okay, you may think I'm just a hater at this point. And honestly, I'm not. I wish Cathy and Windsor all the best, but this is just...bad.

The last blow...

They essentially built an SDK used by no one, and a group of VCs gave them $11M to keep growing it. Wow.

I mean, seriously look at their NPM metrics: 28 downloads in the last week.

And when looking deeper, most of it was actually generated using stainless (an SDK code generator for APIs)! They didn't even bother to update the default readme for their packages!

I get using tooling like Stainless to build the SDK. But to not have any users? And get 11 million dollars to turn it into a scaleup? That's just mind boggling to me. How would they even make money?

Alright Nyuma, so what?

Who wouldn't say yes to $11M to build a next-generation AI SDK?

Well remember. The problem isn't that these founders are bad nor got 11M. The problem is that setting the standard for what the future looks lile. People forget that 1 million impressions is a lot no matter what you're doing. They

Pimping out highschoolers

TPOT were going crazy due to this tweet:

This young lady was hired by the Instinct (ex-cluely) company that is claiming to be the best way to make money online. Essentially, what they're doing; a Whop 2.0 - having impressionable people create content for companies in exchange for money.

It's not just them being hired. It's about everything this represents. Chief of Staff? Without a developed frontal lobe? It's shameful how a freshly 18-year old is being pimped by these guys to make thirst traps and push their startup into the spotlight.

Trendfollowers

And the ripples were instant:

There's more to this, but I'll save it for another time. Overall, it's just sad. Sad to see this happening, and sad to see people getting taken advantage of. Linda at the end there was definitely joking, but still - it's not a good trend to set.

It feels like a big section of the industry (Silicon Valley) is just a bunch of grifters and con artists. Not even hiding behind the vision of say the WeWorks of the world; it's all just blatant, get rich quick energy. I'm not even that old, but I just hate to see this. The weird thing is, I think it's generational, and only going to get worse.

Younger Gen Z know that attention runs the world nowadays, especially growing up on social media. So they think that what I like to call "Mr. Beastifying everything" is the way to go. It's not. It's not sustainable. It's not healthy. It's not ethical. It's probably not even fun. It's just a bunch of dopamine hits.

I reeeaaalllyyy hope this is just a bubble and it will pop soon, but genuinely, I don't think so.

Something's coming. And I'm not sure what it is. But I know it's not good.