正在加载视频...

视频加载失败

“Code is becoming management" Former Google CEO Eric Schmidt says programming has already changed more in 6 months than in the last 20 years. Top engineers are no longer writing most of the code themselves. They’re managing fleets of AI agents with objectives, verification, and long-running tasks. The leverage...

16,707 次观看 • 5 个月前 •via X (Twitter)

23 条评论

Michał Piszczek 的头像
Michał Piszczek5 个月前

Wrong frame. Engineers aren't becoming managers. They're becoming the accountability layer for systems they can't fully predict or inspect. That's a different job - and a different failure mode.

Jon Hernandez 的头像
Jon Hernandez5 个月前

Source:

Pierangelo Raiola 的头像
Pierangelo Raiola5 个月前

"Code is becoming management" is the headline. The real shift, the bottleneck moved from syntax to taste. Anyone can ship a feature. Very few can decide which feature deserves to exist. Distribution and judgment get rerated. Implementation gets commoditized.

Aris Chronos 的头像
Aris Chronos5 个月前

Managing fleets of AI agents. Sounds a lot like what middle managers have done for decades, just with different subordinates.

Anon Ymous 的头像
Anon Ymous5 个月前

Former Google CEO Eric Schmidt doesn't know shit about software engineering.

Creao AI 的头像
Creao AI5 个月前

It's not just that engineers are managing agents, teams that actually operate this way are shipping at a pace that changes the unit economics of building entirely. Different headcount math, different deployment rhythm. Our example: 99% of production code written by AI, 3-8 deploys a day, team of 25

Since AI 的头像
Since AI5 个月前

😎

Harry Tandy 的头像
Harry Tandy5 个月前

the terminal is basically turning into an autonomous corporate org chart

NameFave 的头像
NameFave5 个月前

Managing AI agents is the new coding. fits that shift perfectly.

Thinkonomix 的头像
Thinkonomix5 个月前

Managing agents is still management. The real question is what happens when the agents start managing each other.

Benjamin Brown 的头像
Benjamin Brown5 个月前

no that was already how engineering worked

InteractiveST 的头像
InteractiveST5 个月前

So code's gonna undercut construction budgets and cause buildings to collapse on people? Impressive coding.

Thomas Gatliff 的头像
Thomas Gatliff5 个月前

Sounds like a guy trying to push a narrative. "no longer writing most of the code themselves." he is correct, but they are just productivity tools. Not a "Im just a manger" role. AI LLMS are simply the newest CAD tools replacing draftsman. They are not replacing software dev

Tetragrammaton 的头像
Tetragrammaton5 个月前

I have no idea why anyone listens to Schmidt.

Raman Bose 的头像
Raman Bose5 个月前

for what products people are writing codes?

Amaar 的头像
Amaar5 个月前

Absolutely

Kago Kagichiri 的头像
Kago Kagichiri5 个月前

It’s never over. The best expression of language to the agent harnesses will be code. It’s poetry.

Æ 的头像
Æ5 个月前

Your job is overrrrrr

Home 的头像
Home5 个月前

The geopolitical subtext is as important as the economic one. If only a few regions can finance AIs cale infrastructure, then compute becomes concentrated not just technically but financially. That creates a strategic asymmetry where access to capital markets effectively determines access to frontier intelligence systems, regardless of scientific capability or hardware availability globally.

Leah Freiman 的头像
Leah Freiman5 个月前

Agree. The shift isn’t “no more coding,” it’s from writing code to orchestrating systems of AI agents with clear goals, constraints, and verification.

Mari 的头像
Mari5 个月前

We live in an era where the idea matters more than the code. Agents handle the routine, freeing humans to create.

John Allen Berger 的头像
John Allen Berger5 个月前

Wrong grasshopper, the code of bots and agents begins with human software developers but quickly pivots when self learning and self reinforcing entities take over the entire Huang 5 layer stack. Multi dimensional attack by the machine. Think the Matrix.

AI Money Map 的头像
AI Money Map5 个月前

AI expanded my personal capability. I used to do Java backend only — now I write across languages and work full-stack.

相关视频