Apple Engineer Exits for AI Infra Role After Internal Transfer Blocked

The interesting data point here is how this engineer had a TON of success with cold applications thanks to Apple being the last several years of experience. And this dev is “AI-pilled”, just taught it all themselves, built stuff on the side etc.

Not many takeaways beyond yeah, Apple doesn’t seem to be doing too much AI-related work, or if they do they hide it

@GergelyOrosz imagine having the talent in-house and making them job hunt to use it. classic bigco.

@GergelyOrosz I dislike Apple, but I'm actually still bullish on them.

When the AI hype settles, we'll discover we still need things other than AI-s, and cool devices with a ton of cool apps will be one of them.

@spiral_archtct They use Claude Code, this dev was a v early adopter. Just did not find opportunities to build agentic systems. They two are v different!

@GergelyOrosz They definitely have AI adoption internally. One of their CLAUDE.md files leaked recently.

@GergelyOrosz Can't blame them. AI's moving too fast to wait a year for an internal transfer.

@thepanta82 Whenever I learn how they build software I am bearish. Buts and pieces of info and it’s a head scratcher. I’m talking pre-AI sensible practices (lack of in pockets I hear of)

How they build hardware: bullish, totally

@GergelyOrosz wild that switching companies is now easier than switching teams

@GergelyOrosz blocked internally at a company building neural engine chips, then two ai infra offers externally. headcount allocation is the actual constraint.

@GergelyOrosz 大厂内部转岗慢，AI 基建公司反而给了更直接的成长路径。真正想做前沿方向的人，最后还是会被更高密度的工程环境吸走。

@GergelyOrosz same tbh

but instead I left the company I worked at and started my own 😁

@GergelyOrosz I guess the question is, will anyone wean off Apple ecosystem and their premium hardware due to deteriorating software quality. And I just don't see it, not unless things get a lot worse.

@GergelyOrosz Were posts about them being addicted to Claude Code false?

@thepanta82 @GergelyOrosz They have the 2 most important things: 1. A gigantic install base of excellent inference hardware. 2. The reputation as a company that can handle your privacy correctly.

Nobody else has this.

@GergelyOrosz couldn't get a transfer internally so they transferred externally. flawless system.

@GergelyOrosz What gets me is that a motivated engineer literally couldn't move to the team they wanted, so the market got them instead. Companies lose their most determined people this way and call it attrition.

btw, one thing that I'd love to know and always wondered about it - how they decide and how/if they test stuff out with some test audience. I never cared about SD card slot, but there's huge enough volume of DSLR folks that made Apple revert the SD port removal. And I wonder about Liquid glass - either there's some mega revolution is displays coming and iPhones will be transparent and it's just a framework that absolutely had to be done, or they basically didn't do much of any user testing. And I'm curious about some simple stuff like here and there - bugs that are in their DB unresolved forever...

@GergelyOrosz He would be proud of himself for taking the leap

@GergelyOrosz Oh fair point, they are very different!

@GergelyOrosz This reflects how competitive the AI talent market has become. Engineers increasingly want to work where the fastest progress is happening, and if they can't access those opportunities internally, many are willing to change companies to stay close to the frontier.