An X exchange between indie founder Pieter Levels and researcher Herbie Bradley spotlights a narrow but pointed disagreement: Levels claims AI has made building software so trivial that he has swapped paid SaaS tools for quick custom versions and sees talent drifting toward hardware, while Bradley counters that only the act of writing code has become cheap and that reliable enterprise products still demand real engineering.
I don't know if it's obvious information or not but if you talk to random people in San Francisco the general thing they say is that software is commoditized cause so easy to make anything with AI fast (like how I cancelled all my SaaS subscriptions and just vibe coded a replacement for free) and that everyone smart is getting into hardware cause it's still difficult to enter, kinda related to the Midjourney Medical thing too
Bradley’s reply keeps the focus on production realities—security, scale, maintenance—that casual AI prompts have not yet erased, leaving open how many teams will still pay for those layers once basic code generation is free.
Levels ties the commoditization claim to an anecdotal shift of “smart” people toward hardware, yet no hiring data or head-count trends back the observation, so the predicted move remains an unmeasured hypothesis for now.
Positive users express excitement about AI commoditizing software and shifting talent to hardware, while negative users call the claim wrong or overly simplistic and dismiss it with sarcasm or hostility.
No Digg Deeper questions have been answered for this story yet.
China is amazing in niche hardware. Extremely high quality. China is making a fortune selling niche hardware to the west.
They have the ecosystem of parts that we don’t have in the west.
A guy can start an hardware business by putting 100 parts together from different vendors.
@levelsio There's lots of difficult software left to build
@levelsio But can you really vibe code a replacement for something like Figma or Salesforce? 😅
Not sure where the line is.. It's only easy to replace simple utils, not complex systems.
@levelsio Is it "difficult" or just capital-intensive? Hardware challenges might be a bit more physics-based, but they're just technical challenges. But they cost so much more to even experiment with.
I wonder for how long that will still be true...
@xdrewmiko @levelsio This. Anything sufficiently complex that it would require a full-time employee just to create, maintain and expand it is still viable.
@levelsio I think that’s herd thinking, yeah ai will commoditise software, but hardware too, just a bit later. The only competitive advantage is working on things you enjoy, whatever that might be, and doing it better than everyone else. But that has always been the case.
@levelsio the fun thing waiting on the other side of the hardware game is supply chain and manufacturing, and im not sure software heads are ready for that one. not impossible, but not a walk in the park. extremely entrenched leaders in that space
@levelsio All about moats
@levelsio Millions upon millions of consumers won't ever vibe code replacement software though.
@levelsio
@levelsio they do say that, but they're kind of wrong? code is commoditized, software is not otherwise every enterprise would be vibecoding their saas products
I don't know if it's obvious information or not but if you talk to random people in San Francisco the general thing they say is that software is commoditized cause so easy to make anything with AI fast (like how I cancelled all my SaaS subscriptions and just vibe coded a replacement for free) and that everyone smart is getting into hardware cause it's still difficult to enter, kinda related to the Midjourney Medical thing too
@levelsio levelsio air fryer, when?
@levelsio I’m building a recycled GPU tester just to FAFO
execution is getting so much easier that it's causing all of the easy, obvious, normie ideas to get swallowed up and executed.
applies in hardware too!
genuinely exciting for the state of the world and for solving real human problems
but also a bit terrifying for me as a solo builder
@levelsio i'm curious - what would you be building if you were a vc backed founder in sf today?
@levelsio I do see a shit ton of software getting replaced entirely in the years to come. But most people still don't understand AI. They think it's just the ChatGPT chat, they have no concept of vibe coding. Hardware is definitely the better place to be though
@levelsio Isn't most hardware pretty straightforward?
Like most hardware isn't that advanced at all?
Like sure, building cutting edge robotics is a frontier field.
But most hardware isn't frontier work.
@levelsio A lot of the big winners from before were those who got control over a standard. There are fewer and fewer cases where I feel like settling on external software for anything in the first place when I can just hack together a streamlined custom solution in hours.
@levelsio If midjourney has no room in software what am I doing here
@levelsio Absolutely. Not enough supply to fully meet AI demand in terms of hardware.
Id guess a solid 18 months to take advantage of this, stocks, early investing etc. If MidJourney is just the beginning, I’m very excited for the medical/health world.