Sotto Voce

“AI” and “The Cloud” are both hot topics, but couldn’t be more different. AI is new, unproven, and surrounded by hyperbole, whereas “The Cloud” is older, established, and broadly accepted. But online, criticism is mounting against both, not so much for the technology itself but for its misuse. Instead of waiting for big tech to wake up, we can do something about this.

The AI Pitch

AI is the new kid on the block, and too many fanboys are freaking out about what it may do. Of course, “AI” is far too grandiose a term; this isn’t “Artificial Intelligence” but an LLM or Large language model. It’s a technology that predicts text with no understanding of whether it is ‘true’ or not. It’s mostly appears to be true, to be fair, but it has no concept of truth, just statistical probabilities that the ‘next word is likely’. Unlike the fanboys, I’m not freaking out over LLMs for the simple reason that I worked in AI in the 1980s at Intellicorp when the first rule-based expert systems were built. Today the EXACT same predictions and lazy extrapolations were made back. Totally different tech, I understand, but seeing the same gushing over-reaction was so predictable. If you don’t know your history, you really are doomed to repeat it.

I want to be clear that I’m actually hopeful that LLMs will be useful. I’m just against this breathless enthusiasm over-selling it, much like crypto was over hyped. Many skeptical voices are speaking up. There is a constant drip of negative stories from two directions. The first is against the tech itself, citing issues with bias in these models, its energy use, or its inability to actually plan. The second type of story is around its poor adoption, with short-sighted companies betting the farm on badly functioning LLMs and laying off entire teams.

The Cloud Pitch

Unlike AI, “The Cloud” is hardly new. It’s an established technology that has evolved over decades, and is so far from sensational that it’s practically boring. The pitch goes that everything should be in the Cloud: your photos, documents, and passwords. Basically your entire digital life. Doing this is remarkably helpful: all your devices can see all your stuff! Amazing! In addition, it’s all backed up. What’s not to like?

Well, it too has a constant drip of news stories about security breaches, sudden service shutdowns that lose customers’ data, or a customer support mix-up that locks people out of their family photo collection permanently. To cap things off, many companies are using the excuse of “moving to the cloud” to increase what they charge users by moving to an annual subscription model.

It’s the execution stupid

These two technologies couldn’t be more different, but they do at least have the potential to work extremely well together. By putting your entire life in the cloud, it’s possible to apply LLMs to all of it. Running services on your data actually and I’m personally excited by the potential. The problem isn’t with the raw technology but with its execution.

AI and The Cloud are both capable of boosting productivity and, believe it or not, protecting user privacy but unfortunately, that’s not even close to happening. Both of these technologies are viewed through a simplistic corporate lens and used primarily to either slash costs (Cory Doctorow makes a compelling case LLMs should be increasing prices, not reducing them) or coerce users into paying higher monthly subscriptions. These two technologies, together, are laying bare many corporations’ anti-consumer shift. Even Economists are starting to wake up on this issue, understanding that capitalism is more about power than efficiency.

Sotto Voce

A gestalt is emerging: a growing public distrust of these technologies. Not because they are inherently bad but because companies are using them so myopically. Business pressures are “forcing” companies to make decisions that are counter to people’s immediate needs and those people are starting to notice. I see this bubbling up on social media and news stories. A Sotto Voce, a quiet whisper with deep urgency. The risk of course is that if this public undercurrent breaks out into the mainstream, it will overreact and lash out, likely calling for the tech to be banned. Public opinion isn’t, unfortunately, very nuanced.

What do we make of this? Is this just the disgruntled mumblings of a bunch of idealists? Isn’t it horribly naive to push back on corporate interests? Whenever I talk about this with corporate types I get some variation of  “But how are you going to make payroll?” This reply is distressing, tone deaf, and shockingly uncreative. Of course, there is a way to do this properly, we just have to start with the right premise: a consumer’s bill of rights. I’m not saying money isn’t important, but let’s not sell our soul in the process.

Solution 1: Stick vs Carrot

While government action can be hamfisted, I’m encouraged by the EU’s stance on data privacy. Even if flawed, it may be the only way to force consumer rights to triumph over unencumbered capitalism. It’s the stick that one hopes will wake up forward-thinking companies. 

But it would be so much better, and frankly more efficient, if organizations would get ahead of this and offer these services in a way that is sustainable and privacy-protecting on their own. Instead of waiting for the government’s stick, companies could get out in front with a corporate carrot that protects both their customers and their bottom line.

In the 1980s, Apple defied corporate expectations by making design its core focus. This was seen as expensive, a bit naive, and wasteful by many companies at the time. But Apple’s bet created a brand that was worth so much more. It saw a bit further into the future and created a “feature” that only it could provide. That type of commitment was visionary.

Nothing is stopping a company from doing that today with cloud services and correct use of LLMs. Sure, it may not squeeze every possible penny from the brand, but that’s short term thinking. Like Apple, it could create something worth so much more in the long run. This is clearly aspirational and I don’t expect many companies to take this route as it is risky and unproven. But isn’t taking risks what VCs are supposed to be all about?

Note: Please don’t confuse the Apple of 1980 with the Apple of today. It pays lip service to privacy and competition and is a completely different topic I won’t get into here. Just suffice to say, Apple is VERY short term focused and I don’t think they will be offering a carrot any time soon….

Solution 2: The Long Term Play

A more future-looking solution is built around the inevitable growth in processing power and mass storage. This could unlock a local, open-source server running both cloud and LLM solutions for consumer use. This is clearly a stretch as most people don’t want to manage their own hardware. But if the Innovator’s Dilemma has taught us anything, it’s to pay attention to ‘clunky solutions’. As the services and interoperability stabilize, one small shift could change everything.

The same software could be virtualized and run in a cloud-based cooperative but with the right governance model to engender confidence. Just look at how Mastodon has turned into a self-maintaining system with no IPOs in sight. Mastodon may not be as big as Twitter, but it is still extremely large and it is paying for itself, thank you very much.

This isn’t just a change in software and hardware, but in business models as well. It tries to take the need for constant profits out of the system but leaves just enough money to actually pay for the infrastructure.  It’s just so crazy it might work.

Our Kobayashi Maru moment

Today’s status quo is what VCs are trying to do today. Solution #1 is what these same VCs could do if they had any vision. Solution #2 is obviously longer term as open-source software usually lags corporate software. But that lag will shrink over time. In addition, the sotto voce isn’t just coming from consumers, it’s coming from programmers and designers as well. More and more of us are working on open solutions and it’s only going to grow. 

This is a long-term play. It will continue inexorably, and I expect it to asymptotically approach what corporate solutions will be capable of doing, especially as hardware costs continue to fall. Ultimately, I hope both solutions come to pass, as it would be a win-win for everyone. We can calm the sotto voce; we just have to listen and get ahead of it. If the game is unwinnable, maybe it’s time to change the game.

Leave a Reply

Your email address will not be published. Required fields are marked *