Readit NewsLots of this article relates to the reasons startups died when cash was freely available - both from VCs and from the markets you were trying to find product in. For example, if you started an online learning company in March 2020, you'd have hit product right away (along with a thousand competitors), and been lathered with cash from every direction. But three years later, all of those startups were struggling, and I don't know of _any_ that survived. That's not a case of the business owners in 1000 discrete companies giving up. That's the entire world economy reverting back to in-person learning, and the disappearance of the ultra-low interest rates for the company to fall back on while it pivots.
In 2025, founders need to be acutely aware of exogenous factors, as they can be business-obliterating events without the social safety net of 0-1% IR.
The other companies working at that scale have all sensibly split off into geographical regions & product verticals with redundancy & it's rare that "absolutely all of AWS everywhere is offline". This is two total global outages in as many weeks from Cloudflare, and a third "mostly global outage" the week before.
So the human is there to decide which job is economically productive to take on. The AI is there to execute the day-to-day tasks involved in the job.
It’s symbiotic. The human doesn’t labour unnecessarily. The AI has some avenue of productive output & revenue generating opportunity for OpenAI/Anthropic/whoever.
The uk gov development service reliably implements huge systems over and over again, and those systems go out to tens of millions from day 1. As a rule of thumb, the parts of the uk govt digital suite that suck are the parts the development service haven’t been assigned to yet.
The Swift banking org launches reliable features to hundreds of millions of users.
There’s honestly loads of instances of organisations reliably implementing robust and scalable software without starting with tens of users.
With a human, you give them feedback or advice and generally by the 2nd or 3rd time the same kind of thing happens they can figure it out and improve. With an LLM, you have to specifically setup a convoluted (and potentially financially and electrical power expensive) system in order to provide MANY MORE examples of how to improve via fine tuning or other training actions.