Ever since Sonnet 3.7 came out, I have no doubt that the software engineer job (and thus, the job market) will change. I sincerely believe that in two years, tops, there will be startups with a valuation of over $100 million, where the majority of the code base (if not all) will be written by AI under the supervision of senior developers.

(I believe developers will still exist, but the capabilities of each of us will grow orders of magnitude, and we will all become tech leads and PMs — I want to write about this separately)

But the last week has been nothing short of explosive. In just one week: Anthropic released Opus 4.1, which is allegedly better at legacy code, and promised more powerful models in the coming weeks. OpenAI published two great open-source models, and there are rumours about GPT-5 coming out on Friday (which raises a question: how good must GPT-5’s performance be that you give away to open-source the models that can compete with your own SOTA?!). Google released Gemini 2.5 Deep Think and the groundbreaking (for video) Genie 3. And just before that, JetBrains announced the development of a new programming language for writing AI code.

Here we must run as fast as we can, just to stay in place. And if you wish to go anywhere, you must run twice as fast as that. How is one supposed to keep up with this? 🏃‍♀️