Produced by PodRescale · 26 Dec 2025Hosted by Lenny Rachitsky

Why humans are AI’s biggest bottleneck (and what’s coming in 2026) | Alexander Embiricos (OpenAI Codex Product Lead)

The content pack PodRescale made from this episode of Lenny's Podcast: 15 finished assets, all public.

What if the biggest thing holding AI back isn’t the tech—but us? In this lively episode, OpenAI Codex Product Lead Alexander Embiricos peels back the curtain on what it’s really like to build the world’s most advanced coding agent.

1:25:13Original feed

What the host didPublished one episode1:25:13 of audio, on their own feed. That is the entire input.

What we didA content packWe transcribed the recording and generated the writing and assets below.

What came out

15finished assets2channels5planned posts

Browse the generated output, check its source and review it before publishing.

In the pack:Show notes8 quotesNewsletterTranscript

Everything this episode became.

One 1:25:13 episode, 15 finished assets. Open any of them.

01Show notes64 words
02Chapters8 · timestamped
03Quote cards8 cards
04Newsletter3 subject lines
05Episode titles12 title options
06Summaries2 versions
07Transcript18,469 words

A sample publishing week.

5 sends across 2 channels, starting the morning the episode goes out.

Example schedule

MonEpisode out
Tue
Wed
Thu
Fri

After launch week4 more assets fill quiet days until the next episode starts its own week.

Normally this is four handoffs: editor, clipper, copywriter, whoever owns the scheduler. For a connected show, the queue runs itself once the host approves it.

Built from a public feed, so nothing was posted.

Check the writing against the source.

Each asset carries the moment it came from, so the host can check it, not trust it.

The current underappreciated limiting factor is literally human typing speed or human multitasking speed.

LennyLenny's Podcast · 1:28

TranscriptLenny's Podcast · 1:25:13Play this moment

1:28Lenny

The current underappreciated limiting factor is literally human typing speed or human multitasking speed .

Matched word for word against the episode’s own transcript.

Read it for yourself.

Read the show notes, quotes and selected writing from this episode. Review generated text against the original before using it.

Show notes · as generated

Alexander Embiricos, Product Lead for OpenAI Codex, joins the show to discuss how humans are now the primary bottleneck for AI progress, what it's like to build at OpenAI, and what the future holds for AI agents by 2026. From Codex's explosive growth to its impact on app development and code review, this episode dives deep into the real-world evolution of AI-powered coding tools.

Lenny introduces Alexander Embiricos, highlighting his background and the significance of Codex as OpenAI’s coding agent. Codex is described as a smart intern for software engineering, increasingly capable of writing and reviewing code, and even managing its own training and infrastructure.

Alexander explains how Codex enabled the Sora Android app team to build a new app in just 18 days, with a public launch 10 days later, demonstrating Codex’s capacity to accelerate software delivery.

AI coding agents like Codex are rapidly accelerating software development, but human limitations—like typing and multitasking speed—are now the biggest constraints to further progress.

The current underappreciated limiting factor is literally human typing speed or human multitasking speed.

Pull quote · Lenny · from 1:28

One of the most mind blowing examples of acceleration: the Sora Android app. We built it in 18 days and then 10 days later, so 28 days total, we went to the public.

Pull quote · Lenny · from 0:22

Newsletter example

How OpenAI Codex is redefining what’s possible in tech

If you’ve ever felt like you’re moving fast, wait until you hear how OpenAI operates. In this week’s episode, I sat down with Alexander Embiricos, OpenAI Codex’s Product Lead, to unpack what it actually takes to build at the frontier of AI.

We talked about what Codex is doing right now—catching configuration errors, running its own training, and even helping launch a brand new Android app in just 28 days. But what really struck me? Alexander’s honesty about the real bottleneck in AI: us, the humans. “The current underappreciated limiting factor is literally human typing speed or human multitasking speed,” he told me.

We also got into the culture at OpenAI, where decisions are radically bottoms-up and the pace is relentless—even compared to startups. If you want to know what it’s like inside the room where the next generation of tech is being built, this episode is for you.

Ready to see what’s really driving AI’s future?

Listen now

The episode belongs to