I was musing yesterday more about the idea of an Age of Protocols — @0xstark compared Protocol Theory to Information Theory — the latter inspiring/enabling our current Information Age and potentially Protocol Theory leading to a similar Protocol Age.
One thing that needs fleshed out in that comparison is what are the equivalents in Protocols to e.g. the internet, html, desktop publishing tools, etc. which made creating and distributing information vastly cheaper. How will creating/distributing/running protocols get vastly cheaper and/or more powerful?
I don’t think either Brownlee or I particularly need AI, or, to put it another way, are overly threatened by it. Yes, ChatGPT would have written several thousands words far more quickly than the hours it took me to write this Article, but I am (perhaps foolishly) confident that they would not be original and valuable enough to take away my audience; I think it’s the same for Brownlee.
The connection between us and AI, though, is precisely the fact that we haven’t needed it: the nature of media is such that we could already create text and video on our own, and take advantage of the Internet to — at least in the case of Brownlee — deliver finishing blows to $230 million startups.
How many industries, though, are not media, in that they still need a team to implement the vision of one person? How many apps or services are there that haven’t been built, not because one person can’t imagine them or create them in their mind, but because they haven’t had the resources or team or coordination capabilities to actually ship them?
This gets at the vector through which AI impacts the world above and beyond cost savings in customer support, or whatever other obvious low-hanging fruit there may be: as the ability of large language models to understand and execute complex commands — with deterministic computing as needed — increases, so too does the potential power of the sovereign individual telling AI what to do. The Internet removed the necessity — and inherent defensibility — of complex cost structures for media; AI has the potential to do the same for a far greater host of industries.
I think the internet age upsides were all just Moore’s Law and had little to do with the higher-level stack. Things got cheaper, faster, and frictionless because computing got way cheaper. If there is a protocol age, I suspect it will be driven by the tightening constraints of the post-Moore age. I think we have maybe another 100x improvement on Moore’s law but it’s by no means certain, due to both physics uncertainties and geopolitics type issues. So assume a plateau after another couple of doublings, say just touching the picometer range past 1nm. We now have more of a zero-sum playing field with different dynamics. I suspect the key features of the bounding box will be:
Faster-faster-faster will give way to multi-temporality (ie improvement rate curves will decouple)
More cozyweb, less public internet
“ZIRPy” outbreaks of abundance dynamics that are very local rather than global
Scarcity that is mediated by non-market or post-market things, like sumptuary culture. I made an analogy earlier to NFT-gating culture and medieval sumptuary culture.
If you consider the possibilities of this bounding box, I think we don’t get a “vastly cheaper and/or more powerful” as a broad cultural condition. Instead you get “abundance enclosures” within a general condition of scarcity. Those enclosures are likely to be “puddles of agentic AI” that join human society. Protocols will provide the scaffolding for managing this condition without collapsing into a post-institutional crisis. It’ll be a high-tech version of the high middle ages…
Hmm. Wondering if there’s a really low-level way to enchippen protocols so they get maximal benefit from what’s left of Moore juice. In a way IoT is partly a “protocolize Moore’s law” vision but not quite.
So question: If you could do a hardware accelerated “Protocol chip” what would it do?
One obvious candidate is hardware accelerated zk circuits (which already smell like Verilog to me). Encryption obviously.
What else? I think R2D2 is the true protocol droid, able to connect to anything. How would you enchippen that? Superficial example: ESP32 is protocolized Arduino since it adds wifi.
If we can define this as a grand challenge it would be interesting to get semiconductor people thinking about it.
Right now, our gatekeeper style permissions checking naturally means there are chokepoints in handing from system to system. Capabilities models, like @expede’s invention of UCAN, and the work that Spritely and others are doing, mean that you can move between people, systems, agents, orgs and anything in between, without having to pre-negotiate a particular kind of auth.
Today, we see this INSIDE closed systems, like all of the services inside the boundary of AWS.
Once you have an “open IAM” system – that has to be built on something that is cryptographically secure and capability based – you open up safer remixing of all sorts of systems.
There is a lot of talk of new “OS” layers. These are the systems that are closest to having “capability” systems and we get some “interop through the file system”. Anything that claims to be an “OS” will need some of these properties.
Google spent years working on Fuchsia as a next-gen capability-based OS, but I believe they fired all the people working on it now.
Bonus thought: it’s generally been infeasible to think about N-factorial bespoke interop. There have been some articles about two AI systems figuring out encrypted communication channel on the fly. So my agent and your agent just come up with custom comms as needed. Is that the “death of protocols”?
I’m also interested in what tech is enabling an Age of Protocols. Another question worth asking might be which technologies are necessitating protocols. So protocols as a reaction.
For example, maybe the explosion of data is just driving everyone nuts so we’re reacting with protocols around phone use, fact checking, etc. Higher mobility among certain groups is resulting in a need for governance protocols at popup cities.
I think a flaw in what I’m suggesting is that infotech isn’t directly hazardous, relative to other tech that has caused protocol-reactions in the past
I don’t see this as the death of protocols. Protocols are sociotechnical in nature, right? So if the technical capabilities for ad hoc, bespoke agent-agent comms are solved then it just opens up a world of possibility for novel social interactions and relationships.
The death of parts of the technical process of protocol design.