Last year, there was a marked change in the programming capability of the latest Large Language Models (LLMs). AI-generated code became professionally usable, and in combination with the release of tools such as Claude Code, Agentic Engineering was set to take off in the work place.
How has it played out?
Ideation, rate of iteration, initial design, debugging and Mean Time To Recovery (MTTR) have all improved. Since Thariq Shihipar's tweet on the effectiveness of HTML outputs from these tools, I've been approaching problems differently too, reducing decisions to multiple-choice, then often refining these choices toward an optimal solution.
One of the wonderful things about LLMs is that they do not get bored. Where I fatigue two hours into fixing a sticky bug, AI just keeps going. It's immensely satisfying, too, turning round a task in an afternoon that previously would have taken me a week.
However, it's messy. Engineers are handing out 10,000 line PRs like they're going out of business, the most successful SaaS providers are grappling with demand, and Junior Engineers are not being given the opportunities to learn. One METR study highlights a bizarre phenomenon where experienced engineers believed they had completed tasks faster using AI tools, when in fact doing so had slowed them down. Clearly, it's sub-optimal to reach for these tools as your default.
Do we even need engineers?
Like many others, I wondered whether I'd be in a job following the rollout of coding Agents. Past the initial shock, using these tools in anger has offered reassurance.
When prescribing instructions to an agent within the bounds of a topic and tooling which I understand deeply, I produce high quality output at great pace. However, as soon as I dip my toe into knowledge that I don't have, or an Agent starts building with tools I haven't used, I work up a sweat. I become wholly reliant on the understanding of an LLM which hallucinates at least 1.8% of the time. And if this poorly understood code needs debugging? Best hope I don't run out of tokens!
Yes, we need engineers. In this new world where lines of code are almost free, quality is your only differentiator. It's engineers who dictate that quality. In five years, these tools will undoubtedly have evolved; I am optimistic that those who place their human-in-the-loop most effectively will be the ones who come out on top.
The Great Extrapolation
"Cancer is cured, the economy grows at 10% a year, the budget is balanced — and 20% of people don't have jobs."
~ Dario Amodei, CEO of Anthropic, describing a very possible scenario (May 2025)
Claude is brilliant. I use the models everyday. This is not a hit piece on AI, nor Dario specifically; These tools are altering the landscape of Software Engineering, and they deserve to be.
However, I find these wild claims from Silicon Valley CEOs exhausting. If a doctor stated "we've got much better at treating breast cancer over the last five years", and then used that fact to claim "thus, we will likely have a cure for all cancers in 1-2 years", you would scoff. Let's please apply that same, measured scrutiny to such provocative claims. AGI is not coming in one to two years. No matter how many times its definition is changed.
"How do you plan to achieve this cure for all cancers, doctor? On what timescale can you build all these data centres? Will you require another $500B?"
~ a rational human, replying to this hypothetical doctor's ludicrous declarations
Closing notes
I feel for these Silicon Valley CEOs. In spite of delivering incredible technology, they're deep in an economic bubble; Dare they voice realistic claims or honest timelines, and we'll all feel the repercussions of a dumping stock market.
Today remains an exciting time to be an Engineer. There's so much enthusiasm around tapping into the potential of these tools. If you have any interest in Software Engineering, take some comfort in that. Lean into it. Make these people's ideas a reality - one that's secure, performant, and intuitive. And finish your prompt with "make no mistakes".