https://whilstinarrakis.wordpress.com/feed/atom

The new Dutch CGT that just passed is going to kill all their locally incorporated deep tech startups, because they take longer to ramp up. They might grow revenues but will take so long to reach profitability that they’ll fall under its jurisdiction (30m in revenues, 5 years old).

This is so idiotic that idiots should be offended by being compared to it.

But you know what? I welcome these acts of economic self-sabotage from people who don’t understand that wealth comes from entrepreneurship, because the fallout from this will be for everyone to see. So the next idiot who gets voted in by lazy idiots and proposes some sort of tax on unrealized capital gains will have to face the countless case studies that measured its impact on the Dutch/Norwegian/etc ecosystems.

Few additional notes:

• If you really want to tax entrepreneurs, make it easier to go public so they’re no longer “unrealized”—heck, make it HARDER to stay private even if you really don’t like it. The problem is that applying this at a cap of 30m is so incompetent I don’t even know where to start.
• Private companies doing a deal at a valuation cap of $1bn ARE NOT WORTH $1BN. It’s a contract with the shareholders that if the company were to be sold at that cap, or above or below, there’s a nice agreed way to split the proceeds. But until it’s on the public market, that price is a fiction. Other countries also seem not to appreciate this part (Another Norther European country uses the cap on SAFEs as the implied price to be discounted to measure one’s valuation for tax purposes—presumably if one raises uncapped, the value of the company is then infinite?).
• Even if the law doesn’t pass, the damage is done. I cannot imagine a single investor who would not force the reincorporation of a Dutch company (and to some extent already happened because their system is stuck in the past with notaries and wet ink signatures).
• I feel strongly about this, because it could have been me. In term of assets, I have very little. But I have started a company that is old enough, and could get funding via grants, that would have fallen afoul of this law and cause someone like me to be liable for tens or even hundreds of millions of euros, despite having equity that could be worth effectively zero.
• The thing that in all this grinds me the most is that the Dutch government did not even want to pass this law, but they were forced by the ECHR, and because it’s a coalition government they could not agree on another solution. Two examples of why Europe is a lost cause.

There is always this period in a revolution where it is not quite clear when it starts.

The Industrial Revolution is the obvious example. You can point to the steam engine, or to pumps, or to textile machines, or to Watt, or to the first factories, and depending on what you pick you get a different start date. Before some inventions you are clearly not in it, and after a cluster of inventions you clearly are, but in between there is this long grey zone. People have spent entire academic careers trying to decide where to draw that line.

World War II is almost the opposite. The invasion of Poland very clearly started it (again I am sure someone will argue that it started at Versailles or even earlier). That part is not fuzzy at all. But then for months almost nothing happened (unless, of course, you were Polish). So the beginning is sharp, and what it means in practice takes time to unfold. Different kind of grey area.

AGI has felt like some mix of both.

For decades there was not even a clean definition of what AI meant, let alone AGI. And for a very long time, really since the 1950s, the core argument was symbolic versus connectionist, or something neurosymbolic, and which of those was supposed to be the foundation. That argument ran for generations.

Then transformers came out and started to scale.

Then ChatGPT was released.

That was the first time people were really talking to something that felt like a thinking machine. But it was still a kind of oracle. You could ask it things, it would answer, but it could not do anything. It could not act in the world. So calling it a stochastic parrot was not unfair. It was just reproducing text.

So the debate kept going. When would AGI really arrive. What kind of breakthrough would be needed.

Then, over the last weeks, there was Claude Code, then Claude Cowork, then OpenClaw ClawdBot (now MoltBot). These things appeared extremely fast. And with OpenClaw something shifted, because suddenly a non-expert could talk to a machine and it would do something. It did not require a wizard [1]. It would write programs by itself. It would run them. It would spawn other processes.

That was when it started to feel different.

Looking back, even though it has only been weeks, it feels like we crossed something. AGI happened when machines were able to write programs from human intent. You can now give a desire to a machine and it can turn that into executable steps. It can break it down. Each step can spawn other things. Some things it cannot do directly, so it writes a program that emails a person and says, can you do this, we will pay you this amount. And the person does it. But the loop is still controlled by the machine.

That feels like a thinking machine that makes other thinking machines.

AGI is the moment when a machine can take intent and turn it into self-extending machinery.

Some people will nitpick this. They will say it has to be fully autonomous, that it has to independently deploy everything, that it has to do all of it end to end. Planning, world modeling, intention, yadda yadda yadda. Whatever. Just academic arguments against the inevitability of the pebble that is causing the avalanche in front of our eyes. Because at this point there is already recursiveness in the system. The machine can build parts of itself and the workflows around itself. That is the thing that matters.

AGI does not mean omniscient.

You still gotta paid bills and still go to work. Space-age post-scarcity communism is not quite around the corner.

We do not have machines with their own intentions. We are also constrained by compute. We do not have enough GPUs. We don’t have enough energy. The takeover stories are not what is happening. But the barrier feels crossed. We now have something that can build components that improve the system it is part of.

From here there are still a lot of things left to invent. Full embodiment with robotics. Machines that can act in the physical world. Machines with something like consciousness. Companions in your head. None of that is done. The steam engine did not mean the Industrial Revolution was finished, it meant it had begun. This feels the same. We are going to stay busy inventing for a long time.

In a few centuries there will probably be a chapter in a history book about this whole arc. It will start with census machines in the 1890s, then go through scaling laws, then nuclear energy and Trinity as an accelerant, then the Intel 4004, then neural networks, then transformers, then ChatGPT, and then the moment when machines could write programs.

And that moment will sit inside that long curve.

After that, the story changes. It is no longer just about humans inventing machines. For a while it becomes humans and machines inventing together, and then eventually machines inventing on their own.

Yesterday, walking through central London towards the Tube in the evening, one of those absent-minded walks where you just let your brain run, it hit me that something very large had happened and it had sort of slipped in without anyone really noticing.

We crossed it.

[1] https://www.computer-dictionary-online.org/definitions-w/wizard

(This is a short clip entirely made by AI using Hailou)

Hollywood is just a few Python scripts and a bunch of GPUs away from being transformed. We’ll be able to take a comic and generate an entire movie from it. The quality will be extremely poor for a few months, but then it will become perfect.

You might not like comics? That’s fine – think about any books, poems, or literally any piece of media being used as a prompt.

Hollywood will fight this tooth and nail, then eventually give up and become essentially a giant law firm, suing platforms to extract royalty payments.

The only people affected will be those who want to create movies the traditional way.

You don’t have to watch Marvel movies all the time. The point is that those who want to will be able to, and more importantly, people who want to create will suddenly not need a $200M budget to make a decent movie.

Go and look at the first few videos posted on YouTube almost 20 years ago, and now compared it to what has become an indispensable tool for daily life, and possibly one the greatest educational tools humanity has. The evolution of generative AI movies will have a similar pattern over a shorter timespan.

(I think this might be the first video ever uploaded on YT)

With this endless offering, content discovery and exploration will become more crucial. For a while, the cost and time to generate movies will mean there’s going to be some centralization in discovery and distribution, plus the prompts at least in the short term will be what drives the black box of neural nets.

This is inevitable – the only question is the timeline.

Google hit peak advertising. They are way above $300bn ARR, at this point their cap on growth is  

• How much time people spend on a browser and search engine 

• How much time people spend on Android 

So presumably they saturadted all the cheap growth and user acquisition channels. What remains now is 

• They degrade their search engine quality and force people onto their properties, acting as middlemen for transactions and essentially charging endpoints for traffic that they used to deliver for free (already happening, interesting antitrust implications) 

• They develop new hardware which allows them to reach new audiences (e.g. AR glasses) and create more screen time to deliver ads (already happening) 

• They render the UX on Safari for iPhone rather clunky, and they move people onto their own app, and over time it becomes a super app where they exert more and more control over a user (this is 2025) 

• They create additional revenue streams from their moonshot factory. Everything there has failed with one big exception, Waymo, which could revolutionize both Level 5 autonomous driving and fractional ownership of cars. Americans spend $1,200 or so on mobile per year, vs a minimum of $10,000 on cars. It’s huge! (This is developing, but because of the physical and legislative component of this, it will take longer to scale) 

• They go all in on AI and try to change how their advertising works. And this is the part where I want to focus. 

Everything I listed, more or less works in the same way. There is an ad inventory, the ad gets displayed, Google gets paid. With AI the whole ad would have to be baked into an AI assistant to recommend a product, or when a product get recommended without conflict, then it needs to figure out how to get an affiliate link (but that long term is not enough, it really needs to manage as much as possible of the transaction including the payment and ideally the money lending to the user) 

I think in the short-term Google revenues will peak, possibly Gemini will start getting adoption, Google profit will skyrocket, briefly, like RIM and Nokia did because of cutting R&D to hide the fact that their core business is eroding. 

At this point there is a future where Gemini (or whatever DeepMind are going to call their new AI agents) becomes an indispensable partner for the everyday life of a billion of consumers, and this generates so much economic activity that not only makes up for the lost ad losses, but also outstrips it. I think at that point DeepMind will not be spun out as a separate company as they tried multiple times over the last decade, but actually Google will become a subsidiary of DeepMind, existing only to manage their legacy ad display infrastructure, and maybe the front end of sales. 

The more AI becomes woven into our lives, the more important DeepMind (and OpenAI, and Anthropic, which somehow fulfil similar roles at Microsoft and Amazon) will become. 

The “battle” for AI is not just agents but is a battle for most of the future economic growth. 

The UK should not host investor meetings begging for investment, because the UK already represents the greatest macro investment opportunity available. If the UK were to match the U.S. in terms of wealth per capita, it would be a $2.3 trillion opportunity.

As a friend keeps saying, being poor is a policy choice, and there is no reason why the UK can’t be the most prosperous country in the world. We don’t even need to go that far – we just need to match the average U.S. GDP per capita and it would be a tremendous market opportunity, similar to the opening of markets like China or Eastern Europe, just without the Communists (well, let’s see how the week goes first).

Sure, there are bigger markets today, and it’s easy to deploy money there. But try to get the money back from China – good luck with that. Or navigate the courts of India. Or expect not to have political interference in Vietnam, a country with a one-party system. So suddenly the UK looks stable, the institutions are strong, it’s part of the Anglosphere, part of Five Eyes, etc. Basically, it’s the perfect place to deploy money at a macro scale.

What’s missing? A government that really implements the best policies for growth, even when they clash with their manifestos or the wishes of the electorate. Because the NHS needs reform, because excluding the military, everything can be run way more efficiently by the private sector.

We need to copy the U.S. as much as possible, not because they are great in every domain, but because they are better where it counts – i.e., economic growth.

They have massive advantages, but let’s face it: we are not exactly Somaliland, so the UK has zero excuses to be in the situation we are in. London is relatively wealthy, but the rest of the country is poor (and this is the inexcusable part) and it is getting poorer. Blaming Hitler, Attlee, and Truss is to some extent correct, but they left office in 1945, 1951, and 2022, albeit in slightly different circumstances.

It’s 2024, so which way are we choosing?

I am deeply embedded in AI progress, keeping tabs on both research and product advancements and learning as much as I can. And still I get surprised. Recently had to complete an undisclosed task, so I went and looked how the top companies in the world did it, trying to copy their structure, learn lessons etc. Then used an LLM and the output was so much better than anyone else I had seen.

This was not an example of cheap stuff done quickly at scale. It was top quality stuff that would have taken experience and sensibility to create, and still would have taken me hours to complete.

We are all saying that AI is going to change everything and so on, but when one crosses that barrier it is just incredible. There was this moment in the 1990s when one could connect to the internet, get email and the web, and suddenly increase business massively while the competitors were not. This is happening again, so really I hope your competitors are big large incumbents with very complex and custom ERPs so that changing processes is difficult. The winners of tomorrow are not those that will just use AI at the endpoints (so that a knowledge worker is augmented), but that will use AI in their corporate process and reinvent it (removing the decision maker in the first place, and then use AI to increase the inputs to a level a human would have not been able to manage).

The sky is the limit once more.

We have entered the endgame of AI progress. Companies should have responded to the DeepMind AlphaGo/Zero by appointing a Chief AI person, and mostly did not. With the Bing/OpenAI announcement we have passed a barrier in the narrative, and your board should empower ASAP a team in the company with the sole and explicit purpose to disrupt your core business using AI as much as they can. Because it’s happening. LLMs are the beginning of the end, and there are several more milestones to go (ongoing learning, understanding, robotics) but for a lot of tasks they are good enough, and more importantly they not just help with existing processes, they allow users to do things that were not possible before, and this is a recipe to mess up your business – and if you think you are immune you still can’t assume that your own customers won’t be.

AIs are going to be the main interface we will use to interact with the world. If you are concerned about the dominance of Google Search or Microsoft Windows you have seen nothing yet.

Sure you can regulate or sue left and right, and maybe this will destroy one company, or give you a nice payout for one quarter. But it doesn’t matter. Like Napster did for music the appetite for this tech is high and so are the user benefits, externalities be damned.

When the history of humanity will be written centuries from now the Singularity will be initiated by the invention of the transistor, but only because it will allow us to build AIs.

The joke used to be “it is called AI when it does not work, because when it works it is called something else” and no-one is making this joke anymore.

Got sick again this week (yay) so self-medicated with massive amount of YouTube, and spent quite a bit of time watching one of my favorite subjects from my childhood: laserdisc games. In the process I unearthed some stuff that did not know about Dragon’s Lair:

• The game was developed by three companies. RDI was trying to develop a console (the Halcyon), but the unit economics were so horrendous that they pivoted into making an arcade game.
• Don Bluth Productions was recovering from a financial failure. Their first animated feature, The Rats of NIMH, underperformed at the box office (Bluth said that the kids were either watching ET or playing at the arcade instead of watching his movie), so they explored making video games and partnered with RDI.
• The third company involved in making Dragon’s Lair was a publisher, Cinematronics, which had just went through a Chapter 11 restructuring the year before. With this background it is a miracle that anything shipped.
• There are a number of unused scenes in the Dragon’s Lair disc, and it suggests the game had some limited form of choice of movement within the castle (branching paths were only added in Space Ace, Cinematronic’s following laserdisc game).

Here Dirk moves a log and reveals a trapdoor that leads to an exit, instead of taking the right door like in the normal level.

In the released game an electrified cage surrounds Dirk and one can only move forward. But here it suggests the player can choose between three paths, possibly the lava men level at the left, chewing door front, and an unused level to the right

The cave to the right takes Dirk to the unused level and it features a tentacle monster and some jumping

EDIT: Another hidden level is present on the disc, featuring three one-eyed fleshy pig orcs (porks?) throwing spears at Dirk. The rough coloring (I think there are no special effects or complex coloring palettes), plus the gameplay resembling a traditional game with very repetitive animation to me suggests this was an early gameplay test, possibly the first playable level created to show to investors before exploring a more cinematic approach to gameplay that dominates Dragon’s Lair.

In the released game Dirk also approaches the potions but is attacked by a green slime monster. In the deleted scene it suggests that the potions have each a use and the player can choose which one to pick: one is neutral, one buffs Dirk (an extra life perhaps?), and one is poured over his sword adding a flame effect – similar to the endgame sword that kills the dragon, maybe suggesting some sort of proto-inventory system (which was later expanded in Dragon’s Lair 2 with hidden objects).

EDIT: A sequel was developed, Dragon’s Lair 2: Time Warp, but sat unused until 1990 because of a lack of an agreement between Don Bluth Productions and Cinematronics. That one also had a missing level (the Pirate Ship) but it was never fully developed and instead was reused in the absolutely awful PC game Dragon’s Lair III.

EDIT: also found that RDI worked on an original prototype of a storytelling console called Shadoan in 1979, which used a roll of printing paper and a keyboard (below). It can be seen at the 4:12 mark on this collection of interviews about the making of DL. The Shadoan evolved from paper into a film projector before morphing into a proto-Halcyon, but the toy industry was not interested. You can see the obvious Dungeons & Dragons influences on the Shadoan, as well as the re-use of the wizard’s den in Dragon’s Lair.

• Anyway just found fascinating to discover something new about a game that I thought I knew like the palm of my hand, almost 40 years later and wanted to share.
• Given the extreme financial difficulties of the companies involved in the production, I think gameplay was cut down as well as anything involving choice would have created an interruption in a scene which would have been irritating (normally breaks are only between levels or when Dirk dies, and IIRC it could take ~1 second for the laserdisc player to find the new track to play).
• I remember reading about the Halcyon in 1986 or so, as RDI kept pitching the idea of a console with a laserdisc player, which wasn’t a daft idea per se, and the problem was not even the unit economics of the games which pretty much required making a small animated movie (at this time outsourcing to ultracheap Korean animation was starting to get available), just that the gameplay lacked variety (it could only do what we call today quick time events) plus had limited replayability.
• The pricetag for the Halcyon was going to be $2,500 ($6,800 in today’s prices), be voice controlled using a Z80 microprocessor, and it never saw the light of the day but there are few copies that are known to exist as prototypes. It would have beaten the PlayStation by 10 years but I guess the advancement in computing power (meaning it was capable of doing 3d) plus the CD player (which unlike the laserdisc could support digital video and data) is what ultimately won the market.
• Apparently Cassandra Peterson (Elvira) was an investor and owns one.
• Side note: RDI went bust and the guy became a realtor, Don Bluth Productions struggled and then relocated its animation studio to Ireland due to generous tax breaks, and is partly responsible for bootstrapping the Irish animation industry and could be why there is one of the best animation studios, Cartoon Saloon, in Kilkenny (of Song of the Sea and Book of Kells fame). Cinemation went through multiple acquisitions and now they offer gambling apps.
• I have read that an interactive laserdisc adaptation of The Lord of the Rings was also planned (by a Japanese company IIRC but can’t find the name) but unclear if it ever entered production and if it was just going to reuse part of the Bakshi 1978 movie.

(explore this prompt on Lexica)

One of the earliest [1] incarnations of finding stuff on the web was a directory service called Yahoo which was a glorified yellow pages directory [2] that was human curated. While in the beginning this was a good way to navigate the web for new users, very quickly the rate of growth of content became so steep that human curation was not longer the best way to experience the internet. [3]

In the meantime the first web crawlers (programs that would read a webpage and index its text, which is something that parallelizes very well) started to appear, either Sun SPARC servers, on Alpha DEC, but mostly the impact of the internet was underestimated and these companies used search engines as use cases to sell to corporates, instead of understanding that the web was *the* market. So their product was designed not for the customers that were using it. [4]

Linux started getting better [5], and computer clusters based on it became cheaper, based on custom built PCs, so web crawlers were able to have bigger and bigger indexes, and caching more web content. This was the state of the art of search engines. The ~28th search engine to hit the market, Google, was much better than the other, I personally used it because it had the best archive for programming and technical stuff (all the others just returned garbage because they were looking for natural language and became confused where you were searching for code with weird characters and operators). Google took off, more or less became a monopolist globally and here we are still.

Since then some attempt to challenge Google happened either with bigger indexes (Cuil, whose archive – to their dismay – was mostly porn I believe), semantic search (True Knowledge and Powerset, later integrated in Microsoft Bing), privacy (DuckDuckGo, which is neither a search engine nor private, ironically) and more recently You.com (which I am currently using) which allows users to rank results based on their preferences.

But really Google was the last search engine to enter the market. A market that has grown from a few million users to a few billion ones and that defines human culture.

What I am trying to say is that the way search engines work is not because it’s the best way to search, but because of the best technology available in the late 1990s and we are stuck in that paradigm. Maybe there is a better way, but change is hard.

Search engine design, both mobile and desktop, is stuck in a local maxima but the content we consume has changed. A lot of it is in graph form (social networks), data streams (social feed), video content (Youtube and TikTok), ecommerce, and authoritative knowledge (Wikipedia), apps etc. I think a lot of it found that local centralization was more efficient.

I think Google cannot be challenged in their core territory. I mean it’s not my opinion but more a fact; it’s what has happened for 24 years.

[Machine Learning has entered the chat]

I think that the advancements in Generative AI present a point of disruption for multiple industries, but specifically they are a way to break Google’s hegemony on search engines.

Instead of using a big database and searching on it, we need to use that database as training data, and generate results with a neural network. 

EDIT: one of the reasons why this is a big deal is that trained models are really tiny compared to the training data. Stable Diffusion is like ~2 gigabytes but the training data is 100 terabytes (with an estimated training cost of ~$600k for 256 A100s running for 150k hours), and the consensus is that the size of the model could be reduced by one or two orders of magnitude frther still, possibly more. So if you think that the internet is several petabytes, it’s not unreasonable that it could be reduced to a 100 gigabyte model, which is somewhat portable (a 10Gbit/sec fibre connection soon could not be out of the ordinary in wealthy urban areas). So one can see why one could have their own search engine installed locally and not having to rely on Google. The computing cost of running the model would be negligible, as the burden would be on training (which opens another can of worm as we do not have continuous training yet, the model would have to be retrained and redownloaded in its entirety again every time there is new updated info, so there is a great deal of technology that needs to be invented before Google can be replaced, but as others have suggested there might be a hybrid solution in the meantime – to some extent Google already tries to generate some query results when asked some questions and over time they have tried to reduce the traffic to non-Alphabet properties).

Instead of searching for something, and then opening the first few results and scanning for the content we want, while fighting millions of popups, ads, and weight loss scams, one needs to be able to generate the answer they are looking for.

This would also change distribution as installing the entire Google archive on your laptop is not feasible, but a ML model can definitely run (again my money would be on Stability.AI, as this is something DeepMind would be prevented from pursuing [6])

I have been thinking about this post for a while, and in the meantime one search engine already launched (not quite what I am proposing, but it’s using language models to generate URLs https://metaphor.systems/ ). Also Adept is working on an AI assistant which is very impressive, even if (IMHO) they have not nailed the use case yet. Lexica is a search engine for prompts which is also interesting (they start with art, but doesn’t mean that they end up there).

This would completely bypass the distribution monopoly and advertising business of incumbents, as training a model is expensive (for now) but running it is not (the marginal cost of a Google search is negligible, but the cost of maintaining their infrastructure is very much not).

And I am not arguing that this should be only about text. We have the ability to generate multimedia content, so in the near term you could generate a synthetic video on how to replace a faucet or how to solve a calculus problem.

Generative AI is very exciting, and as we backport it to old industries we might find that the incumbent is no longer needed.

—

Notes:

[1] yes I know there was stuff before Yahoo but hypergrowth on the internet happened with the web and allowing commerce on it, and all the other engines (Lycos, InfoSeek etc) have been dead for a very long time

[2] no idea when they (if?) they stopped producing them but it was a printed local phone book of all the businesses in the area, organized by trade, and it was the main way to acquire customers in the telephone age. In the early ‘90s as the cost of printing kept going down these things were huge in terms of pagecount, but in 5 years or so they became ridiculously small until they disappeared

[3] Yahoo pivoted multiple times and it’s still around, if you want to experience what it looked like you can check yahoo.co.jp (does not work if you are from Europe, but you can image search it to see screenshots), which was licensed to SoftBank decades ago and barely updated since

[4] AltaVista was the most interesting example, as they briefly controlled the market but because in their mind the product was used to index corporate documents inside a private network, they never built any kind of spam protection and very very quickly the top results were just spam garbage and everyone moved to Google. Their demise was lightning fast 

[5] I first used Linux in 1995 and although infuriating at times, it was very promising. By 1998 when Google started it was quite good even for enterprise production, and most definitely better than Windows for servers while Unix was fragmented by different vendors making sure their products were as incompatible as possible which doomed them as developers did not bother

[6]  which IMO should be the primary reason why they should do it as a priority, but I guess one is incentivized to forget about the Innovator’s Dilemma when they are being measured by quarterly performance

After yesterday’s MSFT/ATVI deal announcement several people are dunking on Kotick’s NBC comments on the metaverse, and the whole concept in general. Now leaving aside the allegations against him that will hopefully will be addressed soon, he is somewhat right. The game industry has been at a forefront of several technologies (GPUs, networking, UI design, AI and I’m sure there are a dozen more), it remains behind compared to the tech industry on the one metric that matters: who has root on the computer upon which stuff runs.

Apart from two console manufacturers (one of which – Sony – is not a serious contender in the software market), the fight is moving onto data centers, mobile devices and the “metaverse” [1] will be just the name of the content running on some infrastructure.

And whoever owns that infra will ask for rent.

Correctly, game companies are realizing they are just targets for M&A and having scale benefits the offsetting of seasonality of game releases (a game can take 5-7 years from concept to release, with up to 500 full time employees doing nothing else).
Incidentally this is why Epic is kicking up a stink against Apple. Epic doesn’t have a hardware strategy (for now?) and if they don’t have an open distribution platform, all their profits are going to be taken away by the platform holders. Same goes for Snap making AR glasses, and Meta investing on VR.
I suspect the M&A feeding frenzy will continue this year. The streaming platform are also moving into this space (Apple Arcade, Netflix), while traditional media and entertainment falls further behind. It’s a massive market that is only getting bigger.

There has never been a better time to start a game company.

[1] We use the term metaverse for lack of a better word. Before the internet or the web, we used terms like “information superhighway” which also meant fuck all in terms of clarity.