Is AI going to take our jobs?


Hi frieds,

The question posed in the title is likely the one that keeps many developers and engineering leaders up at night these days.

The short answer is no.

I recently came across two fascinating blog posts about AI’s impact on software development.
The first was from the renowned “The Pragmatic Engineer” blog, and the second from Sequoia Capital, a major venture capital firm with investments in giants like Google, Apple, and Oracle.

Here’s a brief rundown of the key insights:

  1. Ask any CEO what’s their biggest spend and they’ll say engineering expenses.
    Leaders have been aiming to reduce these costs since the 1960s. This trend continues today, though the circumstances are slightly different. That’s why we see the same narrative pushed around AI, not for nothing, but the level of hype is somewhat inflated
  2. The emerging solution isn’t about artificial general intelligence (AGI) taking over all jobs. Instead, it’s about increasingly sophisticated models that will be great assistants. Sequoia’s vision is particularly bold, aiming to expand the developer community to a billion users.
    It’s an ambitious goal, yet it represents an optimistic view of technology’s role in our future.

It appears that developers who hesitate to adopt AI might fall behind.

This is a widely held belief—debatable, yes, but generally accepted.

With that perspective, here are four ways I use AI to enhance my productivity and speed:

  1. General AI Usage:
    I use GPT-4 for a variety of tasks including creating images, editing text, and generating code.
    Yes, this service is paid, but it’s incredibly versatile!
  2. Local Code Assistance:
    For my side projects, I’ve integrated Codeium into Neovim.
    This setup helps me with smart code completions right where I need them.
  3. Interactive AI in the Shell:
    I’ve also set up an interactive AI chat in my shell using a tool from GitHub (you can find it here).
    This lets me talk to any AI model I choose during my coding sessions.
  4. Local models run free!
    Lastly, if you’re looking to access all these benefits for free, Ollama offers an easy way to make every model portable and locally accessible. For finding models, Hugging Face is an excellent resource.

Before you go, a quick tip for Mac users:
Raycast offers plugins in its store that integrate OpenAI’s GPT, Ollama, and many other tools, allowing you to converse with AI directly from your OS spotlight!

Thanks for reading, as always - feel free to reply to this email with feedback and questions!

Have a great weekend.

Whenever you’re ready, here’s how I can help you:

ESPRESSO FRIDAYS

Every once in a while I send hand picked things I've learned. Kind of like your filter to the tech internet. No spam, I promise!

Read more from ESPRESSO FRIDAYS

Stop Using Elasticsearch. Use Postgres Instead This issue is brought to you by: Postgres for sensor and machine data Factories, wells, grids, and fleets run on live machine data. Tiger Data keeps it performant and scalable in one Postgres. Get started with Text Search on TigerData→ Do you still need Elasticsearch? I wanted to test that properly, so I put Postgres in Docker, loaded it with thousands of support tickets, and checked how far the same BM25 relevance search used by dedicated search...

Somehow, Atuin Just Got EVEN Better This issue is brought to you by: Give your AI agents memory Learn how Oracle Agent Memory handles persistent context with custom extraction, hybrid search, filtering, and more control over what your agents remember. Read the developer guide → There aren't many terminal utilities I'd take with me anywhere. But if I had to pick just one, Atuin may very well be it. I've been using it for a few years now, and somehow the magic turtle keeps getting better. It...

You’ve Never Seen a Shell Like Xonsh This issue is brought to you by: Redis Iris: Your agents should be getting smarter Unreliable agents fail in production. Redis Iris is a unified, real-time context engine that delivers fresh, relevant context so agents perform at scale. Try for FREE Tell me if this sounds familiar:You start with a quick shell script. Then one step gets annoying, so you call Python. Then Python needs to shell out again. Then you’re parsing strings, juggling quoting, and...