Companies are increasingly adopting AI technologies, with AI-related spending growing by 293% in the past year. This growth is driven by the need for automation, cost savings, and improved decision-making. Non-tech sectors like healthcare and financial services are catching up in AI adoption, with notable increases in spending and usage. Specialised “narrow” AI tools are gaining popularity, and businesses are budgeting for recurring AI expenditures. Despite a slowdown in new AI investments by tech companies, overall adoption rates remain high.
Business Opportunities from the AI Spending Surge:
AI Integration Services: Entrepreneurs can offer services to help businesses integrate AI tools into their existing systems, particularly targeting non-tech sectors that are rapidly adopting AI.
Industry-Specific AI Solutions: Develop and market AI tools tailored to specific industries such as healthcare for automating radiology workflows or financial services for automated financial advice.
AI Training and Support: Establish training programs and support services to help companies and their employees effectively use AI tools, addressing the knowledge gap in non-tech industries.
AI Vendor Partnerships: Create partnerships with popular AI vendors to provide bundled services and solutions, enhancing the value proposition for businesses looking to adopt AI technologies.
AI-Driven Data Analysis: Offer AI-driven data analysis services to help companies make better decisions by leveraging the data they collect, especially in sectors like finance and healthcare where data volumes are high.
These opportunities leverage the growing trend of AI adoption and the need for businesses to efficiently integrate and utilise AI technologies to stay competitive.
Opportunities…until Google, OpenAI or Microsoft eat your lunch, right after you have de-risked a business proposition or proven a market opportunity. The juggernaut will soon squash Otter.ai, Fireflies.ai and our local Tactiq.ai, all in the business of transcribing and summarising online meetings, as Google Meet, Teams, Zoom etc roll-out built in alternatives. Everyone worries about AGI. The real problem is that the major LLM foundries are heralding a “business opportunity singularity”. Every third party AI opportunity is an just an “arbitrage on time” where the window is getting shorter every day.
People are still discovering the wonders of being able to chat to a PDF and there are countless small companies making money from those people even though the SOTA (state of the art) is now free.
The technology is moving fast but the consumer adoption is moving at a snails pace by comparison. I also think AI is suffering from something akin to the last mile problem and it will need many small companies to help solve this. This would be similar to companies today built on helping people use Microsoft word.
I think this is telling of the situation, Replit laying off 20% of their workforce the week of GPT-4o’s release. To survive companies need to be insanely agile and find a place in the market that makes sense for them (if any exist).
It is not consumer adoption that is the relevant metric. It’s employment. I already run a team about 3 times larger than I did a year ago on one major project, but I’ve not taken on a single new human. On another we’ve downsized humans by 30% but expanded by 200%, in part human but not as new employees.
I’m living the “dream”. Only it isn’t, because all my (viable) competitors are doing the same things. And those “consumers” trying to read PDFs with AI are going to wake up one day and ask why they need to even try; they no longer have a job.
Altman has said he anticipates the first “solo unicorn” worth $1B soon. Not great for the economic model everyone assumes in our neoliberal world is “the only way” to generate and share the fruits of production.
I completely agree @Bruce and funny enough we were talking about this exact topic in the AI Guild last night. @ken has a differing view to us and I would really welcome other people’s perspectives on this as to be honest I would really love to be proven wrong!?
I did a bit of digging into this and I agree the writing is on the wall but we aren’t there yet and might not be for a few years at least.
Over the past 6 months job growth in the US has been across the board.
It is only in the past month that we are starting to see a decrease in job growth from the professional services sector and ‘Information’ job growth falling as well is telling. But this is at the very pointy end of the statistics.
If you look at how long it took for technology to remove jobs it takes a while.
Some examples that I think are telling
DVD Rental Clerks
Netflix founded in 1997
Cracks start to appear in Blockbuster 2002
Netflix launches streaming 2007
Blockbuster in serious trouble 2008
Travel Agents
Expedia launches in 1996
Wasn’t until 2008 that significant job losses were felt by traditional travel agencies
Telephone Operators
235,000 telephone operators in 1930
The first direct distance dialling (DDD) service in 1951
1984 down to 40,000 telephone operators employed
2022 down to 4,300 telephone operators employed
Word Processors and Typists
1983 Microsoft Word released
150,000 in 2000
44,000 in 2022
All these examples to me speak to a realistic timeline. Now I completely agree this change will be much more turbulent than any of those because it isn’t one type of job being removed from the economy but all jobs but I still think it is realistic to expect this change to happen over the next 5 years and it’ll be gradual to start with. A lot can happen in 5 years as well and there is plenty of opportunity in between here and there.
I’d point out that I am saying 5 years and not 10 or 15 like a lot of these examples would suggest and the reason for that is Nvidia’s sales figures.
It is obvious that corporates are running toward AI adoption as fast as they possibly can and they have insane amounts of capital to throw at this which has not been true at any other stage in history. That points to this happening much faster then history would have it but even with that pace of change I still don’t see major cracks in the labour market yet.
Yep, i suspect that job will last for about as long as the supervisor to the automated telephone operator system. Once trust increases that person won’t be needed.
Yeah, I think it will last a bit longer. Companies produce financial statements that need to be audited by a third party. I think the same will eventually apply with their AI output. Whether that’s internal or external.
I think that goes back to the saying “the future is here it is just not evenly distributed”. It takes a long long time for some things to get around the world. This might be one of them.
I suspect AI job loss may accelerate faster than these historical examples suggest due exponential increases in capabilities and, massive capital investments in AI by the very few that can. The consolidation in industries across the board will appear slow at first but I think we’ll see a “sudden” impact as companies begin to fail or transform in “emergency mode” leading to faster (existing) job loss than new job creation. Unlike past disruptions, AI impacts multiple (almost all?) industries simultaneously. The only caveat is that the exponential growth in energy requirements to produce this result at scale may reign it in a bit.
As Consultant, you only have to know a little more than your customers than sell them the knowledge gaps to save them time.
As Coach you help them to figure their solutions
Identify thencurrent situation and desired outcomes
Do guided reach with them,
Reflect on the possible avenues
…
It could be a blend as a teacher.
Identifying solutions you tried and like and feel confortable to resell, YouTube is full of these
Right now AI is mostly good for genereating ideas, still you need an expert to review to understand if what is as been generated make sense.
For the jobs market going down, we are maybe bluffed by the post COVID bubbles, plus yeah easy to do 264% better when AI is truly accessible to public since 18 months.
I would like to warn everybody that most of all AI (LLM) offered like ChatGPT, Gemini and … simply want and train on your data while it simply rehearsing what everybody knows or should know.
For now it invent nothing simply, rewrite, put together words that as been pushed in the LLM, this is why OpenAI, Google, Meta and others make them available for free or almost in exchanges of your data to train their next version.
If you don’t believe me, please read the Terms of Use.
Yep this is very true, they most certainly want your data.
while LLMs are undoubtedly informed by their training data, their responses are not mere regurgitations. Instead, they reflect a sophisticated interplay of learned patterns, contextual understanding, and probabilistic generation, resulting in original and contextually appropriate outputs.
The recent work done by Anthropic has put very important evidence to this, showing that LLMs work in a similar way to the human brain by utilising patterns and structures learned from extensive training data to generate contextually relevant and coherent responses, rather than simply regurgitating memorised information.
Microsoft’s $1 billion supercomputer data center now under construction in Mt. Pleasant, WI, is expected to cover 1,030 acres when completed in 2034. The firm is buying raw undeveloped land for the project.
In terms of Microsoft’s plans for a 5GWe data center, and possible co-location of nuclear reactors to power it
I suspect energy will be a significant limiting factor here. If Microsoft and OpenAI are planning to take until 2034 to create the infrastructure needed for their projections we might be quite a way off and that assumes lead-times and costs don’t continue to rise as they have been.
