ZUCK will replace mid level engineers with AI

How credible is this?

I'm going to be honest, Zuck has kind of gone crazy over the last year or so. This is coming from someone who empathizes with him and thinks he's treated unfairly by the press, particularly with politics. I wouldn't treat anything he says nowadays very seriously. As Ilya mentioned, all of these Silicon Valley CEO billionaires have an agenda. For FAANG in their particular, they want to convince everyone of the following:

1. AI will quickly reach a point where it can do everything ever
2. Big Tech will be the driving force behind this AI revolution, making hundreds of billions of $$$
3. (buy Big Tech stonks and drive up their share price)

It's been 2+ years since ChatGPT came out, and it's still struggling to fully independently write code that would meet expectations as a Meta E3 (junior), let alone a Meta E4 (mid-level). I'm sure it'll get there someday, but definitely not within the next 1-2 years.

How to prepare?

There's nothing special you need to do - Just keep developing senior-level behaviors that are independent of the code (i.e. follow the advice in Taro 😛). This is something every engineer everywhere should already be doing, AI revolution or not. Here's some great resources to get started:

• [Course] Grow From Mid-Level To Senior Engineer: L4 To L5
• "What are software engineering fundamentals?"

Here's another fun AI doomsday thread: "The AI frenzy is really making me rethink my career - What direction should I take?"

How credible is it that some algorithm will be able to write code at e4 or even e5 level? That will probably happen. How credible is it that some algorithm will replace e4 engineer in the foreseeable future? Very very unlikely.

Current algorithms are unable to properly reason, they are unable to solve novel problems, they are unable to prioritize. There is nothing yet in the works that overcomes those... not sure deep learning as it is currently is able to do so.

Engineering is not about being a coding monkey but about solving abstract problems where sometimes the real solution is "do nothing".

Silicon Valley leaders have an agenda. What they say about "AI" needs to be looked at through that filter.

Here is the exact quote from Zuckerberg: Meta is "going to have an AI that can effectively be a mid-level engineer at your company that can write code." (reference)

I bolded the part that is most important. The AI can write code, but as Ilya mentions, there is a lot more to being a software engineer than code.

So, yes, I do think AI tools can credibly become as good as, or better than, mid-level engineers at coding. BUT there are other parts of the job that will become more important. (That's why we started Taro.)

My other comment is that tech execs and founders are strongly incentivized to make an outlandish claim because it shows their innovation. This can inflate valuations and lead to hiring better people.

But remember that tech execs are notoriously bad at predicting the future. (see any number of quotes from Elon, Zuck, or Sam Altman)

I think many people, understandably concerned about the potential impact of AI’s rapid advancement on their jobs, are reading too much into this statement. The same applies to other statements made by MANGA-level C-suite executives, whether it’s Jensen Huang, Sam Altman, Jeff Bezos, etc.

I don’t mean to downplay the concerns you and others are likely feeling. It’s completely valid to be anxious about how rapidly this technology is advancing and how it could affect our roles in the near future, especially given recent layoffs and the concern expressed in the framing of your initial post.

That said, it’s important to remember that the primary role of a CEO at any tech company is to promote their company’s value, increase revenue/profits, and boost stock prices for shareholders, who are more often than not VCs or private equity firms who tend to know very little about day to day software engineering. So, like other CEOs in his position, Zuck is likely exaggerating. Let’s not forget that this is the same person who was heavily promoting the Metaverse not too long ago.

We all saw how that turned out.

Now, this isn’t to say there’s no reason for concern.

However, engineers should only be seriously worried about an AI replacing them when we see a demo of an AI system where someone verbally asks an agent, like one that has just been started up, to analyze an enterprise/MANGA-level codebase of at least a few hundred thousand lines, and is able successfully build production-ready features or resolve complex, non-obvious bugs. And all of this would need to happen within a significantly lower time frame and cost than an average engineer could manage, with no issues, and in a consistently reproducible manner. By the way, this is fundamentally impossible with the current architecture of transformer-based models because hallucinations are an inherent limitation of the architecture.

Until then, the real concern will continue to be engineers who know how to effectively use AI taking all the jobs.

https://www.salesforceben.com/salesforce-will-hire-no-more-software-engineers-in-2025-says-marc-benioff/

What should we be making of statements like this?

We’re not adding any more software engineers next year because we have increased the productivity this year with Agentforce and with other AI technology that we’re using for engineering teams by more than 30% – to the point where our engineering velocity is incredible. I can’t believe what we’re achieving in engineering