I work for a $20 billion tech company (not a big tech company, but not a small one either). Our company strongly encourages us to incorporate agentic tools like Claude Code and Codex into our daily routines. Management has made it easy to use Anthropic and OpenAI products with an enterprise subscription. I can sense that our business leaders want us to move more swiftly and be more productive to avoid falling behind in the AI race.

My team has adopted Claude Code extensively, and the number of daily and weekly PRs we have closed has increased significantly. I’ve noticed that we’re also more willing to commit to more projects. My team benefits from most of our code being written in TypeScript. However, some other teams with legacy code bases seem to have a bit of a harder time using these tools compared to us.

One thing that surprises me with a AI is you can have people working on the same code base. Some can be very effective with AI, yet there are others working on the exact same code base who cannot get good results. Some people don’t really seem to be taking the time to get good at writing prompts and plans before having agents execute on them.

Your experiences reflect mine.

My employer has a slack channel where we can share prompts, tips about how to use AI outside of coding, Q&A. I've learned a lot from my coworkers there.

I work for a $50B+ company (is this big tech? idk), but I’ll answer this because we are fully embracing AI (Cursor, Claude Code, cloud agents, AI reviews, you name it)

> Have you noticed faster pace of development?

Yes, our org has had a 50% increase in PRs since Opus 4.5 released.

> Have you seen changes to code quality or code review?

Yes, significantly more bugs (no exact number), but consider it maybe 3-4x in volume. However, nothing catastrophic and everyone just uses AI for fast-follow fixes anyways. The company as a whole is embracing this style of development for better or worse.

> Do teammates that use these tools complete sprint tasks faster than those who don't?

Yes, but my entire team uses them. I’d say the ones who use it more effectively (crazy skill setups, better tooling/commands, better scaffolding) finish much faster. Probably 80% of my team still uses Cursor in the one-shot way with very vague requirements, and don’t have the AI connected to github, jira, slack, etc which can actually feed really important context into decision making.

If I do something more than once a day, I write a custom slash command for it. This has personally 2x’d my pace.

The biggest shift isn't the speed of coding, but the shift in 'seniority' expectations. You're no longer just a writer of code; you're an editor of high-volume output. The mental fatigue has shifted from 'how do I solve this' to 'is this generated solution actually robust for our scale'. Big tech feels more like orchestrating a fleet of junior agents than solo deep-work now.

I work for large tech company, not FAANG.

> Have you noticed faster pace of development?

Yes, and it terrifies me.

> Have you seen changes to code quality or code review?

Yes, because of the volume of PRs, I am sure some people are just approving changes with a quick glance

> Do teammates that use these tools complete sprint tasks faster than those who don't?

Best engineers got better

What terrifies me more like cultural thing:

1. Engineers who push a lot of code are trying to get promoted faster so they won't feel the pain

2. Engineers who genuinely care and trying to land high quality work are getting labeled as "progress blockers" and getting negative outlook, because they won't be promoted, they need to support the slop added by other engineers. And I don't like this state of the culture.

Also engineers who want to get promoted quickly, usually they are very good at talking and can present a tiny skill template as something huge developed over 2 quarters of experience with latest and greatest, while people who actually do the work are more silent and heads down executing, I want them to be promoted but can't impact a lot of things

This reminds me of the pendulum that swings between every company I've worked at

For a time, its all about shipping faster. And rushing things out.

Then there's a major bug/incident/unhappy customer, and management swings back to code quality

I work at a large social media company - it has become a popularity contest on who can ship more AI slop the fastest. Real productivity gains are questionable tbh. Slop is replacing more slop.

I work at Google as a SWE SRE. That means a lot of the code I touch is on tangly complicated projects doing bug fixes and improving things. Not so much greenfield development unless it's a smaller tool change.

AI coding is definitely speeding me up, but not quite in the magic way people at smaller companies might be used to.

As a good code complete it's awesome. As a brain storm buddy, very good. AI code reviews are awesome and very low risk - if it makes a mistake it's usually just being too cautious.

As an end to end agentic coder -pretty bad track record. Giving it smaller tasks? Pretty good, saving me many hours of boilerplate/cruft a week at least.