It seems to me that there is a point in a person’s life, career, and/or seniority, where following the crowd stops being called ‘peer pressure’ and starts being called things like, ‘reacting to market sentiment’.
Admittedly, I am a bit too young1 to remember the dot-com bubble popping. At the time, I was running around my primary school playing field, blissfully unaware of anything outside my little corner of the world. I have a vague memory of a school friend’s Dad losing their job2, but as my Dad worked for the Ministry of Defence, it didn’t really affect me. Actually, I was much more interested in the occasional rolling roadblock flanked by MoD Plod3 than anything even remotely compooterz related.
I say this to admit that I don’t know what it felt like in the run-up to that crash. Neither, really, what it was like in 2008. By that point I’d started at Uni, now more interested in computers than roadblocks, I was still too busy running around.
What I do know is what the industry has felt like since 2010, when I got my first tech job.
Smartphones, and Web2.0 fully usurping “the old internet”. The rise and domination of Facebook, and their purchase of Instagram. Twitter and The Arab Spring. Uber4, Silkroad and the first Bitcoin spike ($21 high!), and how Smart Contracts would change the world. The Gamification of everything. Silicon Valley’s brief flirtation with an abundance mindset. The release of Google Cardboard, Oculus, and the HTC Vive5, and how VR was going to change the world revolutionise global entertainment. The second, third, forth crypto spikes. Various rug pulls and straight up theft. Bitcoin ETFs, the Winklevii. Later, NFTs being a thing, and then very much not. Chatbots going from natural language search of dense knowledge bases, to the solution for every website.
The elevation of Will MacAskill6 from interesting philosophical curiosity, to the billionaires’ favourite guru. The release of GPT-3, the pandemic7 and remote work hiring boom. GPT-4, Claude, DeepSeek, and the positioning of LLMs as proto-AGI, and the ensuing arms race.
This period was full of useful progress.
For all their many, many, faults – Facebook was better social media than what came before. Uber did genuinely improve the experience of ordering taxis.8 Bitcoin, and crypto more broadly, was excellent for bypassing incredibly expensive money transfer systems like TransferWise. This is incredibly useful for migrant workers wanting to safely and reliably send money back home to their families. And, obviously, really useful as a pseudo-anonymous means of payment for a variety of things.
But something changed, and while I can’t really put my finger on when exactly that was, I feel it was around the time that ‘web3’ starting to be thrown around.
To me, Web3 has always felt like a random assortment of technologies, unified as technology-first tools, all looking for problems to solve. For me this is namely Blockchain, cryptocurrencies, VR, and AI. There may be others you include, or don’t.
VR is objectively cool, but severally limited in practical application. Even Apple, with all their billions, haven’t been able to make it useful9. Blockchain is terribly energy inefficient for transactions when done via Proof of Work, while Proof of Stake has serious trust problems for a supposed zero-trust decentralised network. Transaction costs are insanely high for everyday small purchases, leading to the minimal uptake we see in legit merchant contexts. The continued necessity to cash out into a local currency10, often in large increments to minimise the fee ratio, further compounds the problem.
And then we have LLMs.
LLMs are incredible. While I’ve always enjoyed writing, it’s never been something I’ve found particularly easy. Having what is in effect my own personal sub-Editor running in-browser11 helps me better structure my thoughts, and keep my phrasing clear and simple12. It’s not perfect, far from it, but it is very helpful.
The current generation of LLMs are incredibly sophisticated predictive text, and that is great. Friends of mine with dyslexia, or various other neuro-diversities, or normies who just appreciate a hint every now and again, have found it to be a complete game changer.
I’ve also had particularly positive experiences using Claude to help me think around some problems we’ve faced at the foodbank, such as safe loading procedures. It being able to provide direct links to legislation and best practice documentation is awesome. Natural language search is a hard problem, and these tools make significant improvements on what has come before.
A single technology being sufficiently adept at addressing both of these problems is really cool.
But, it’s also very clear that there is a genuine panic in the industry right now.
Across the board, every sector within the industry is trying to shoehorn “AI” in to their products. I’ve been told stories where devs have been directed to replace existing traceable algorithmic decisions for things like financial eligibility, with a GPT call. That 30% of code must be “AI written” by the end of the year. That all customer service enquiries will be answered by an LLM driven chat bot, with no human approval13. And even that user-provided content should be modified, without their knowledge or approval, before being displayed to other users.
The message I take from this is simple – it’s FOMO.
Companies are looking, top down, at what everyone else is claiming, but knowing that they themselves are struggling to make this work. So surely they must be missing something.
What they’ve done so far isn’t something they particularly confident is right for their business, but because their peers are doing it, they must have found the secret sauce.
Every CEO of every FTSE and Fortune business is singling AI out as the gaming changing tech, so it has to be…
Right?
Back in 2000, if I’d’ve done something I didn’t agree with, just because my friends in the playground did, my teacher would have called that peer pressure. She would have reminded me that just because others are doing something, it doesn’t mean I have to.
That desire to fit in, to not miss out, that if others are doing it surely it makes sense to do it too, doesn’t automatically go away when you leave school.
While part of maturing and ageing is learning how to better handle your emotions, fears, and impulses, periods of high stress can resort to us taking shortcuts.
And having shareholders, pension funds, VC, and big equity firms on your neck is rather stressful indeed.
Building sustainable, deep value businesses requires addressing human-level problems. FOMO doesn’t lead to reasoned responses, it instead prioritises reactive, knee-jerk, responses.
So, remember to take a breath, and remind yourself; “What is the problem I want to solve?”.
I promise you the answer isn’t, “I haven’t spent enough GPT credits”.
—
- Although, the increasing grey in my hair and speed I’m approaching 40 may suggest otherwise ↩︎
- Also the first person I knew who had a PC at home – and introduced me to Tomb Raider ↩︎
- From AWE Aldermaston, carrying surprising little, all under a black tarp ↩︎
- I was once asked to provide a cost quote for “the uber of social care”; they didn’t appreciate being sent a link to Uber’s crunchbase page which, iirc, said they’d had over 50bn in funding already. ↩︎
- And punching my friend’s ceiling after getting very disorientated trying to stroke a whale ↩︎
- No relation ↩︎
- No relation… ↩︎
- Now if only they treated their employees correctly ↩︎
- I do feel there is a genuinely viable usage for Augmented Reality – as a HUD for directions, etc – but all concepts seem to be based around forcing yet more advertising in to your eyes, which, well, who actually wants that? – beyond the people funding the development and paying for the ads of course ↩︎
- Or other coins, putting it well above the typical tech literacy of ‘normal’ people ↩︎
- I use LanguageTool, and highly recommend it ↩︎
- Or, at least as best it can dealing with the chaos that is my brain ↩︎
- And this is years after this has already blown up in company’s faces see Air Canada ↩︎