Ray X

Founder @Hi_Livia, Ex-founding member @FlowGPTOfficial, Ex-SWE @tiktok_us, EECS @UCBerkeley

[Ray X @_DarrenX]

Fully agree that macro rates are the main driver here — AI CapEx trend hasn’t changed.
That said, while $AAOI’s $471M H1 2027 guide looks strong, the biggest risk remains extreme customer concentration.
Latest filings show:
• Top 10 customers = 95-98% of revenue
• Single largest customer (e.g. Microsoft) often >30-40% (hit 43.7% in 2024)
Optical transceiver orders are lumpy. If one hyperscaler delays or switches suppliers, revenue can swing hard.
Long-term bullish on AI optics, but $AAOI feels like a high-beta speedboat — best as a small position for those with high risk tolerance. Execution + customer diversification will be key.
What’s your take on the customer concentration risk for AAOI?

[Ray X @_DarrenX]

@OhHelloMillions c to the pineapple #HelloCompetition

[Ray X @_DarrenX]

https://t.co/6BgLheZ7jP

[Ray X @_DarrenX]

@CathyChang00 @UCBerkeley 有眼不识泰山

[Ray X @_DarrenX]

https://t.co/rfFgLqmo3r

[Ray X @_DarrenX]

@OhHelloMillions pineapple and banana #HelloCompetition

[Cathy Chang @CathyChang00]

Graduation unlocked.
Rolling into @UCBerkeley in full Magpie Murders opening sequence.
Go Bears 🐻

[Viktor Seraleev @seraleev]

This is what App Store competition looks like after vibe coding went mainstream

[Ray X @_DarrenX]

Professor Jiang: World War 3 Has Already Begun, Let Me Explain! https://t.co/8ChaxG2GEh via @YouTube
Wow, one of the best 2 hours spent in my life.
Thanks @ProfessorJiang_ you are shaping the reality by a lot yourself.

[Ray X @_DarrenX]

Let's see how Postiz works on managing all my social media accounts for all my apps.

[Ray X @_DarrenX]

shipped a whole new app in 2 hours with Claude design and Claude code. Mind blowing 🤯

[a16z @a16z]

Scrolling is on the decline

More charts: https://t.co/2IoX6EtHav

[Ray X @_DarrenX]

How I use my MacBook now

[Ray X @_DarrenX]

I really think in the future, for a one-person company, all you need to do is just talk and observe. If you have done things other than talking, then you need to refine your agent flow.

[Ray X @_DarrenX]

After running a “solo company” for a while, one thing became very clear to me:

A one-person company and a traditional company don’t actually operate that differently at the core.

The only real difference is:
Before, you hired people, trained them, and managed alignment.
Now, you design agent flows, define workflows, and iterate on reliability.

The real first step is building a solid agent flow system:

So that multiple AI agents can operate like a real team — with persistent memory, clear roles, cross-task coordination, and continuous iteration.

Because even a solo company still requires organizational design.

Two fundamental constraints:

1. Every agent has limited context.
2. The more precisely you define roles and split tasks, the stronger the overall system becomes.

AI is not “one super employee.”
It’s a team — one that still needs to be designed, structured, and managed.

Just like humans.

No one performs well when overloaded with unstructured complexity.

The real leverage is not “having AI.”
It’s whether you can organize AI into a company.

[Ray X @_DarrenX]

@CathyChang00 Sound is too low in the video

[Om Patel @om_patel5]

THIS GUY BUILT A TOOL THAT LETS CLAUDE CODE AUTONOMOUSLY TEST YOUR ENTIRE iOS APP

you point it at a simulator and say "test everything"

Claude navigates the whole app on its own through the accessibility tree and screenshots. it figures out the UI by itself.

it taps buttons, fills forms, opens every screen, tests every feature, and checks every flow

in 8 minutes it found every bug the developer missed

then it checked the debug logs for errors and gave a structured summary of everything it found

no XCUITest scripts, no test maintenance, and no more writing confusing, complicated test cases

one prompt and that's it

[Ray X @_DarrenX]

@OhHelloMillions @ashley98zebra #HelloCompetition