Agreed. On my Twitter timeline lots of people are saying "the world will be completely different in 6-12 months' time". On October 17th, 1980, most accountants' work day still looked the same as a year before. It took decades for spreadsheets to get widespread adoption. I think LLMs are revolutionary, but the (r)evolution will take decades to materialize. Our work days will look completely different... in 2043.
> I think LLMs are revolutionary, but the (r)evolution will take decades to materialize.
1980s was a different world. Today, the distribution that BigTech has, enabled by this thing called the Internet, changes the dynamic completely.
Given the investment, I believe LLMs will change some of the industries within 5 years, simply because it is a new but natural way to interact with Computers; and Google/Apple already put an Internet-connected Computer in everyone's hands.
This LLM hype is very real and is nothing like web3 (which lacked 10x utility over web2, imo).
When low/no-code started its hype? 5-10 years ago? Or something like that. First release dates:
Mendix 2005
Outsystems 2001
Appian 2004 (no info before)
Quickbase 2000
Zoho Creator 2006
~20 years ago. I agree with the idea that it’s a new cool way of interaction with systems, but I also think you trust too much in our development cycles. Five years is nothing. Companies will try and fail, try and pivot, try and barely swim for years. Few of them will arrive alive at stage 2 where the real deal starts.
If it were easy, it would hit frontpages last week, because small companies and single developers can prototype things 10-100x faster than any bigcorp.
The primary issue that I see here is that as a natural language interface to a computer, I don’t know that there’s enough profit to support the hardware and training costs. Also, there are copyright questions involved. There also issues of hallucination, and limiting these neuters the creative aspect of the LLMs.
Personally, I truly hope that this technology gets cheaper and cheaper, and that it can be run on a device the size of a pager in pocket with a pair of AirPods on. If that were my computing environment 90% of the time, I’d be thrilled.
I hate and am terrified of what LLMs will do to society, and I do think they're overhyped, but the revolution won't take years.
They can already fully replace human proofreaders and content marketers. They're very close to being able to replace graphic designers, voice actors, and commercial photographers. I expect we'll see many more unsexy jobs like those being replaced within 5 years.
Throwing out the hot take (for HN) that computer intuition is about to be a much stronger destabilizing force than spreadsheets were. And that’s really where change comes from, IMO - the status quo showing cracks
