# Software Engineering With LLMs in 2025: Reality Check  ## Metadata - Author: [[Gergely Orosz]] - Full Title: Software Engineering With LLMs in 2025: Reality Check - Category: #articles - Summary: AI tools are becoming common in software engineering, especially in big tech and AI startups, but developers remain cautious about their reliability. Executives are more optimistic about AI's impact than many engineers, who often find traditional methods faster and safer. Overall, AI coding tools show promise but still need improvement before they can fully replace human work. - URL: https://newsletter.pragmaticengineer.com/p/software-engineering-with-llms-in-2025 ## Highlights - The reason Google has “custom everything” for its tooling is because the tools are integrated tightly with each other. Among Big Tech, Google has the single best development tooling: everything works with everything else ([View Highlight](https://read.readwise.io/read/01k0841e628j6yyqnnad3g5xs0)) - **Claude Sonnet** is another tool most Amazon SDEs use for any writing-related work. Amazon is a partner to Anthropic, which created these models, and SDEs can access Sonnet models easily – or just spin up their own instance on Bedrock. *While devs could also use the more advanced Opus model, I’m told this model has persistent capacity problems – at least at present.* ([View Highlight](https://read.readwise.io/read/01k0847k702hrj1e1h84bax18d)) - In 2002, Amazon founder and CEO Jeff Bezos introduced an “API mandate.” As former Amazon engineer Steve Yegge [recalled](https://gist.github.com/chitchcock/1281611): > “[Jeff Bezos’] Big Mandate went something along these lines: > > 1. All teams will henceforth expose their data and functionality through service interfaces. > > 2. Teams must communicate with each other through these interfaces. > > 3. There will be no other form of interprocess communication allowed: no direct linking, no direct reads of another team's data store, no shared-memory model, no back-doors whatsoever. The only communication allowed is via service interface calls over the network. (...) > > 6. Anyone who doesn't do this will be fired. ([View Highlight](https://read.readwise.io/read/01k0849ga5kgve7qdpfvskwv5h)) - Since the mid-2000s, Amazon has been an “API-first” company. Every service a team owned offered APIs for any other team to use. ([View Highlight](https://read.readwise.io/read/01k0849z0zn5p4nsj5z1x82b9b)) - One startup asked not to be named because no AI tools have “stuck” for them just yet, and they’re not alone. But there’s pressure to not appear “anti-AI”, especially as theirs is a LLM-based business. ([View Highlight](https://read.readwise.io/read/01k084gbh5evswj848jpjrtysd)) - [](https://substackcdn.com/image/fetch/$s_!zSH4!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F964e5827-9721-4541-9aa2-6cc2aa612230_1222x552.png) Using Linear MCP: sharing learnings with the team ([View Highlight](https://read.readwise.io/read/01k084e1a1txhx88jxvvt7cqzj)) - [](https://substackcdn.com/image/fetch/$s_!CXfq!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffb98d631-0311-4d6f-8708-d4cadb5db62c_1600x512.png) Asking Claude for options: a few things that worked for an engineer ([View Highlight](https://read.readwise.io/read/01k084fd2afrdgqx95bdctxyay)) - **This startup is a reminder that AI tools are not one-size-fits-all.** The company *is* in an unusual niche where ML pipelines are far more common than at most companies, so the software they write will feel more atypical than at a “traditional” software company. ([View Highlight](https://read.readwise.io/read/01k084k25zz9kn361x6zfa3p8y)) - [Armin](https://lucumr.pocoo.org/) is the creator of Flask, a popular Python library, and was the first engineering hire at application monitoring scaleup, Sentry. He has been a developer professionally for 17 years, and was pretty unconvinced by AI tooling, until very recently. ([View Highlight](https://read.readwise.io/read/01k084m72z9tps9zr3mj8jj3wj)) - [Simon](https://simonwillison.net/) has been a developer for 25 years, is the creator of Django, and works as an independent software engineer. ([View Highlight](https://read.readwise.io/read/01k084pn0tjh092knhetzqtv98)) - **“Coding agents are a thing that actually work now**: run an LLM in a loop, let it execute compilers and tests and linters and other tools, give it a goal, and watch it do the work for you. The models’ improvement in the last six months have tipped them over from fun toy demos, to being useful on a daily basis.” ([View Highlight](https://read.readwise.io/read/01k084pysgd0fdtphx81wa5j1z)) - Martin Fowler is Chief Scientist at Thoughworks, author of the book Refactoring, and a co-author of the Agile Manifesto. ([View Highlight](https://read.readwise.io/read/01k084qswnvsq7re3cd305mksn)) - “I think the appearance of LLMs will change software development to a similar degree as the change from assembler to the first high-level programming languages did. ([View Highlight](https://read.readwise.io/read/01k084qwad436kzarpd2j45ze0)) - LLMs are making the same degree of impact as high-level languages made versus the assembler. The distinction is that LLMs are not just raising the level of abstraction, but also forcing us to consider what it means to program with non-deterministic tools.” ([View Highlight](https://read.readwise.io/read/01k084r57zyxrtyncm8rez6jpe)) - #2: How much do devs use AI? Developer intelligence platform startup [DX](https://getdx.com/) recently ran a study with 38,000 participants. It’s still not published, but I got access to it (note: I’m [an investor](https://blog.pragmaticengineer.com/investing/) at DX, and advise them). They asked developers whether they use AI tools at least once a week: • **5 out of 10 devs** use AI tools *weekly* across all companies (50%) • **6 out of 10 devs** use them weekly at “top” companies (62%) ([View Highlight](https://read.readwise.io/read/01k084sankwm4tk8cr2c24krhr)) - The last one is where I pay a lot more attention. For seasoned software engineers: most of these folks had doubts, and were sceptical about AI tools until very recently. Now, most are surprisingly enthusiastic, and see AI dev tools as a step change that will reshape how we do software development. ([View Highlight](https://read.readwise.io/read/01k084wmbvnv4dpg5qmxtqka4v)) - There seems to have been a breakthrough with AI agents like Claude Code in the last few months. Agents that can now “use” the command line to get feedback about suggested changes: and thanks to this addition, they have become much more capable than their predecessors. ([View Highlight](https://read.readwise.io/read/01k084x43rgqkgndjmtyrvbjzg))