v9: Looking at the AI and market trends

Things are gonna get harder before they're gonna get easier 🥲

📍 Brighton, 🇬🇧 

🥃 TL;DR

  • Took a break, back to regular posting

  • Most news is just noise

  • Few words about market trends

  • I’m craving craft and quality code

  • Do less

  • Best advice: Don't chain yourself to programming, keep things simple, don't marry a tech stack

  • Grok Code Fast 1 is a legit Claude alternative

Hey and Happy Monday everyone!

I'm back to my newsletter.

I took some time off. I'M SORRY OKAY!

This will happen again. Not gonna lie and say that it won't.

I just took a bit of time to think about my life, take a break from my usual professional activities, detox (as much as I could). Now feeling refreshed and ready to get back to hustle! 💪

I haven't been bookmarking many links lately. I wanted to make this experiment where IF I am to sit down and recall all the things that happened over the summer, which ones would I remember? Well, here's the list

NONE! I can't remember anything. This made me realise (again!) how important any news are, really. If it's something important, you'd know about it straight away. No need to scroll those news constantly. Everything else is just noise.

Reminds me of this skit

The only thing that "triggered" me was the recent ones: digital IDs and I already wrote a post about it

And ChatGPT releasing the ACP

The last one I liked because I was and still am kinda into e-commerce, and worked with a bunch of headless e-com platforms.

Today, instead of particular news, I'd rather talk about trends.

Agents everywhere

Lots of people have been saying that agents are everywhere and they are! Lots of new agentic frameworks are popping up, Pydantic AI is still young but very promising because of the team behind it. And smolagents is still my favourite because of how nice and simple it is.

LlamaIndex has its own agentic framework, which is not ready yet and not recommended to be used in production.

PS I've lost all the hope for LangChain, LangGraph and all the lang* family of tools and frameworks. Everything they do is overly complicated, unstable and non-dev friendly.

Engineers aren't going anywhere

This article made my day:

Numbers are in. Companies are looking at ROI and realising that they've been lied to. I mean, obviously AI had some effect in some companies on some projects. BUT. This doesn't mean that every company will be like that.

I'm looking forward to seeing our numbers at Sainsbury's after we launch our AI chatbot. I'd love to see how our AI-powered feature would measure up against our goals vs resources being spent on it.

I can't say much because of NDA, but surely I'll be able to share an "opinion" of mine about a thing or two (in November/December, maybe?). 🙃 Stay tuned! 🙌

Market is picking up

This is what several of my mates have been saying. I'm still yet to experience this myself though.

But I think everyone is starting to learn to live in these tough times, finally done with their prioritisation of things and projects and now unfreezing their hires, and looking to start growing slowly again. Which is great, good news! 🔥

RE the AI topic, after several months of being a heavy Cursor, GitHub Copilot and Trae user (yes, I've got them all!) I'm getting exhausted by the amount of AI slop I see daily. I'm craving some craft.

This reminds me of Kerri's post about why quality is rare. Well, now I'm at the stage of the cycle where I want some quality in my life code. I've been through that fast deliver, loads of code, super productive stage.

I don't think I want that anymore.

Because of that, I noticed that I started living for myself a bit, professionally, and ended up following DHH's advice before even reading his post. 😅

There’ll always be more emails in need of reply, more meetings to attend, and more updates to read. A person can fill the entire workweek with these tasks over and over again. But to stay sane and… | David Heinemeier Hansson | 38 comments

There'll always be more emails in need of reply, more meetings to attend, and more updates to read. A person can fill the entire workweek with these tasks over and over again. But to stay sane and sharp, you must pay yourself first by doing the work that actually means something to you. I feel this acutely as someone responsible to employees, customers, followers, and readers. I could do nothing all day but check up on projects, people, and posts, but my brain would quickly check out if it was just doing that. So quite frequently, I just don't. Don't check in, don't check up, and instead dive into the work that checks my own intellectual boxes. Programming for the love of it. Experimenting for the hell of it. Researching for the fun of it. In another age, I might have been tempted to apologize for such privilege, but screw that. Privilege is wonderful. You should do your best to earn more of it. Even if you have to carve it out of the bare rocks around you. Ironically, the best way to do that is also to choose to always pay yourself first, however little at first. By solving your own problems, tickling your own interests, chasing your own curiosity. That's where you'll find the motivation to elevate your talent. To turn interest into competency. And once you've developed some competency, you'll be rewarded with more privilege to build it further. This is the virtuous circle of merit.There'll always be an endless list of work that could be done. You'll never get through it all and onto your own priorities, if you continue to put them at the bottom. | 38 comments on LinkedIn

Speaking of other advice, let me share some of the best advice I've heard/gotten over this summer 👇️ 

  1. Don't be afraid to let go of a programmer you. If you love it - great, keep at it. But it doesn't mean that you have to do it on your job. Your job can be something different (talking about Engineering Management and what not). In fact, on the job, most of the time, we're not doing things that we'd prefer to be doing anyway. It's usually what the company wants. And that rarely matches with what you want.

  2. Simple scales really well. Which is what AI builds are missing most of the time. Most, if not all, AI code is way over-engineered. I usually find myself erasing a good part of it.

  3. You need to stop associating yourself with a tech stack. Tech stack doesn't matter. Especially in today’s AI world. Keep that child-like curiosity alive in you and focus on developing those problem-solving skills.

I think I lied to you about not saving links. I had this one link saved, actually. I felt this was an important one.

But then later Grok Code Fast 1 was released and I've seen what it can do! When my credit finished in Cursor one day, I was forced to switch to another lighter model. This was the time when I decided that Sonnet 4 can be substituted.

Especially at these prices: Grok Code Fast 1 costs $0.20 in and $1.50 out vs Claude's $3 in and $15 out 🤯

Anyway, like I said, I think I'm back to my regular prof activities, including postings. So I'll see you guys next week! 👋 Maybe… 🙃 

🎓️ Top Learning Materials of the Week

Meta Database Engineer Professional - I have been working with databases a lot lately, so I thought I'd become better at it. So I'm going through this curriculum, and I really think its curriculum is great! Firstly, I found SQL for Developers course by Yandex Practicum but I'm not willing to pay £600 for an intro to SQL course, I'm sorry. 🫠

⚒️ Tools of the Week

LobeHub Icons have all the logos of AI companies, which are really in every project lately! 😅

Pydantic AI, as I a said in the post, is one of the most promising agentic framework out there becuause of team behind it, imo. 💯