Robert Dickert

Product-focused software engineer, educator, mountain biker, musician, coffee roaster

[Robert Dickert @rdickert]

This may be accurate for a different reason than you'd expect: Costco checkers don't "pretend" to check your receipt. They count items. They are diligent and exact in this, not lazy. That process is fast and serves as a checksum that catches a variety of possible problems (not just theft, but checkout mistakes including double scans). The person using an AI agent needs to do the same thing: Find ways to review output that catches problems without slowing down for things that are low-stakes or unimportant.

[Matt Pocock @mattpocockuk]

My first AI course is here.

It's the perfect way to go deep into building great AI apps.

We're building a complex, agentic search app from scratch, in 100% TypeScript.

And for the first time, we're running it as an async workshop. I'll be by your side, helping you out.

[Andrew Ng @AndrewYNg]

There’s a new breed of GenAI Application Engineers who can build more-powerful applications faster than was possible before, thanks to generative AI. Individuals who can play this role are highly sought-after by businesses, but the job description is still coming into focus. Let me describe their key skills, as well as the sorts of interview questions I use to identify them.

Skilled GenAI Application Engineers meet two primary criteria: (i) They are able to use the new AI building blocks to quickly build powerful applications. (ii) They are able to use AI assistance to carry out rapid engineering, building software systems in dramatically less time than was possible before. In addition, good product/design instincts are a significant bonus.

AI building blocks. If you own a lot of copies of only a single type of Lego brick, you might be able to build some basic structures. But if you own many types of bricks, you can combine them rapidly to form complex, functional structures. Software frameworks, SDKs, and other such tools are like that. If all you know is how to call a large language model (LLM) API, that's a great start. But if you have a broad range of building block types — such as prompting techniques, agentic frameworks, evals, guardrails, RAG, voice stack, async programming, data extraction, embeddings/vectorDBs, model fine tuning, graphDB usage with LLMs, agentic browser/computer use, MCP, reasoning models, and so on — then you can create much richer combinations of building blocks.

The number of powerful AI building blocks continues to grow rapidly. But as open-source contributors and businesses make more building blocks available, staying on top of what is available helps you keep on expanding what you can build. Even though new building blocks are created, many building blocks from 1 to 2 years ago (such as eval techniques or frameworks for using vectorDBs) are still very relevant today.

AI-assisted coding. AI-assisted coding tools enable developers to be far more productive, and such tools are advancing rapidly. Github Copilot, first announced in 2021 (and made widely available in 2022), pioneered modern code autocompletion. But shortly after, a new breed of AI-enabled IDEs such as Cursor and Windsurf offered much better code-QA and code generation. As LLMs improved, these AI-assisted coding tools that were built on them improved as well.

Now we have highly agentic coding assistants such as OpenAI’s Codex and Anthropic’s Claude Code (which I really enjoy using and find impressive in its ability to write code, test, and debug autonomously for many iterations). In the hands of skilled engineers — who don’t just “vibe code” but deeply understand AI and software architecture fundamentals and can steer a system toward a thoughtfully selected product goal — these tools make it possible to build software with unmatched speed and efficiency.

I find that AI-assisted coding techniques become obsolete much faster than AI building blocks, and techniques from 1 or 2 years ago are far from today's best practices. Part of the reason for this might be that, while AI builders might use dozens (hundreds?) of different building blocks, they aren’t likely to use dozens of different coding assistance tools at once, and so the forces of Darwinian competition are stronger among tools. Given the massive investments in this space by Anthropic, Google, OpenAI, and other players, I expect the frenetic pace of development to continue, but keeping up with the latest developments in AI-assisted coding tools will pay off, since each generation is much better than the last.

Bonus: Product skills. In some companies, engineers are expected to take pixel-perfect drawings of a product, specified in great detail, and write code to implement it. But if a product manager has to specify even the smallest detail, this slows down the team. The shortage of AI product managers exacerbates this problem. I see teams move much faster if GenAI Engineers also have some user empathy as well at basic skill at designing products, so that, given only high-level guidance on what to build (“a user interface that lets users see their profiles and change their passwords”), they can make a lot of decisions themselves and build at least a prototype to iterate from.

When interviewing GenAI Application Engineers, I will usually ask about their mastery of AI building blocks and ability to use AI-assisted coding, and sometimes also their product/design instincts. One additional question I've found highly predictive of their skill is, “How do you keep up with the latest developments in AI?” Because AI is evolving so rapidly, someone with good strategies for keeping up — such as reading The Batch and taking short courses 😃, regular hands-on practice building projects, and having a community to talk to — really does stay ahead of the game.

[Original post: https://t.co/I3alxNs0vn ]

[Robert Dickert @rdickert]

#BSW Is your AI app a product or a feature?

[Robert Dickert @rdickert]

I'm at Your AI Startup Is Just a Feature (And That’s Why It Will Fail) @sched # https://t.co/FTOBIxcx1W

[Robert Dickert @rdickert]

@paulg An AI software engineer. An AI customer support agent. An AI lawyer. I think seeing AI as just doing human roles will morph into whatever the new capability really is - more powerful than a human in some ways but more limited in others.

[The Limiting Factor @LimitingThe]

'Electric Vehicle Fire Myths'

Time to do some FUD fighting! 🤠

Even if you know the stats on EV Fires, there should be some new information in this video for you.

YouTube version linked below.

*Timestamps*

00:00 Introduction
02:21 EV Fire Likelihood
05:09 EV Fire Intensity
10:57 The Triggers of EV and ICE Fires
13:49 Vapour Cloud Explosions (VCEs)
15:13 EVs Are Safer in General
16:26 E-Scooter and E-Bike Risks
18:29 Summary

[Robert Dickert @rdickert]

@KarlDanninger I recently read that The Onion is going to publish to print again. I agree, the fascination with physical board games is very unlike online equivalents. The touch, the weight, even the smell of the paperboard and ink.

[Robert Dickert @rdickert]

@heydave7 @heydave7 Have you talked about how you use LLMs to code? I'd love to hear more. I'm an experienced coder, and I'm interested in how I can (a) make myself more productive by delegating some work; and (b) build tools that facilitate the pipeline that goes from description to code.

[Robert Dickert @rdickert]

@KarlDanninger React & proptypes is probably the hardest place to start. I’m thinking starting with server side code might be the easier/more-rewarding path. Also, it’s worth studying up a bit. I like https://t.co/pWHSsOQ9eE

[Sashko Stubailo 🇺🇦 @stubailo]

I’m working on a software startup to accelerate electrification of commercial vehicles – vans, buses, trucks, etc. 🛻

excited to announce that we’re hiring a founding engineer!

if you know a great product eng who’s looking to get into climate, DM or ✉️ [email protected].

[Robert Dickert @rdickert]

@Dollingerj @TraegerGrills I had the same problem. It turns out my emails were not getting passed through by the system. Customer service people were fine - looks like an IT snafu, at least for me. I got it resolved over the phone - only a manager could see my email. @TraegerGrills please up your IT game.

[Robert Dickert @rdickert]

Just finished @rustlang rustlings exercises and most of the Book with help from Rust By Example: A+ language onboarding experience. This is great work, people. https://t.co/VUPnCZVvkn

[Robert Dickert @rdickert]

@KarlDanninger Awesome. Happy birthday!

[Robert Dickert @rdickert]

My post on #Apollo in Amplify made it into @AWSAmplify Weekly!

[Dennis Hills @dmennis]

Issue 39 of @AWSAmplify Weekly is out: https://t.co/lBH42c8gin with thanks to @Geromekevin, @shawnprzybilla, @mim_armand, @YounesH1989, @rdickert, @dabit3, & @kurtiskemple

[Robert Dickert @rdickert]

I put together this quick tutorial for people who want to play with #Apollo #GraphQL using @AWSAmplify

[OK GROW! @ok_grow]

💥Read our latest blog post - How to set up #Apollo in #Amplify by @rdickert https://t.co/vD4nObBJXK #javascript #graphql

[Robert Dickert @rdickert]

@dan_abramov That’s actually really amazing. Congrats, and thanks for all the hard work learning English

[Robert Dickert @rdickert]

@ecwyne Thanks! Bookmarked

[Robert Dickert @rdickert]

@prisma I find myself a little resistant to code-first as embodied in the example. It feels like I'm reading an AST, not code intended for humans. I do like the colocation of the schema definition and the resolver. Interested to see how composition comes together.