Inhaling Dread

2024-10-08T00:00:00Z

On September 9th, in the height of COVID lockdowns, I woke up and quickly concluded that 2020 had reached climax and the world was ending. Naturally, I had to go into the office that day.

The above photo from that day is untouched, I haven't even fixed the small cropping complaints I have with it, the orange was even more saturated in person. It's probably the best photo I've taken. This was the sky the entire San Francisco Bay area woke up to; our entire world cast in an orange light, unable to see the sun.

I had to bike over two rather large hills to get to my job's soon to be abandoned office to clear my desk's personal items, in a timeslot I had signed up for a month prior to minimize my COVID exposure to colleagues. Of course business MUST proceed as usual, even when every fiber of my being is screaming at me that this is an emergency. Fortunately the air at ground level was, counter to everyone's instincts, pristine. You could do a full cardio routine outside and be fine. The danger was impending, literally looming overhead, suspended by air currents, but not quite touching us yet.

The attitude in the office was... Distraught. Myself and three other colleagues were emptying our desks, picking our way through bins of electronics to find hardware we might've found ourselves needing after six months of work suffusing our homes. What few conversations we had felt like both parties were gently dissociated, everything about our situation seeming so surreal and alien that there was no way it was as close as it felt. The first human contact I'd had in months and we couldn't even find the words to be kind to each other, we were too scared.

On my way out through the lobby I looked south and saw the view above. I won't pretend to have instantly recognized all the meaning and parallels I've found in it about that day and year. I took a minute to line up the photo before getting back on my bike and hauling ass home. The air started to have the occasional hint of smoke in it, I had to book it before the danger reached ground level.

Hellos, Goodbyes, and Promises (Or Lack Thereof)

2024-09-30T00:00:00Z

I'm keeping this short and bittersweet. The death of Cohost is driving me to write, post, and generally interact with the internet in a way that most social media isn't conducive to. So we're resurrecting the dream of the 90s and 00s. Personal blogs baby.

I make no promises about post frequency nor quality. Topics might be anything from video games, computer shit eating my brain, reposting other peoples' thoughtful longform, shitposting, musical uploads, photos I'm proud of... Like, basically whatever suits my fancy. But I do promise to post. And I promise to post things I think I'd enjoy.

This isn't to say I'm abandoning other social media. I kind of like microblogging for small, fun things. I think that's what it's best at (certainly it's terrible for meaningful discussion).

An era has ended. A new one takes its place. We can drag what we loved about our own actions, and what surrounded us, in the latter kicking and screaming into the former.

Eggbug will live on.

The Internet (and The Smartphone) Broke The Capitalists

2023-04-08T00:00:00Z

Author's note: This post originally appeared on Cohost.org more or less exactly as is, including the backdating.

Something that I think people outside the tech sector can't quite wrap their head around is what motivates the trend chasing you see Silicon Valley software companies engage in from time to time.

For a while it was blockchain/crypto/nft stuff. That tanked once it was clear people were just manipulating values for personal gain and the ideas didn't enable anything new (you could already pay cash for drugs and launder money with physical fine art), but it felt like everyone was jumping in trying to figure out some way to involve this cursed math in a product or company, or start a new company around it. Before that it was... probably everything needing to have video? Before that it was social features for everything, because social media (and in particular Facebook) was eating the internet, so Silicon Valley wanted to figure out some way to benefit from these network effects of product stickiness venture capitalists were lauding. I don't remember the late 00s because I was a fucking teenager but I'm sure someone will fill you in and tell you none of this is new, though perhaps it is faster now.

The trend du jour of course, is """AI""", Large Language Models, whatever you want to call Stable Diffusion, and a myriad of other things I will label as Queryable Models or QM because I object to the AI misnomer (though I will use the tag I guess). EVERYONE in the tech sector and their mother's dog is trying to integrate these into their products somehow and I fucking hate it. It's eroding the quality of everything you're going to see on the internet, it's labor hostile, and it prompts the WORST sorts of conversations.

While they definitely want to devalue labor... #

It's really easy to look at the labor hostility point and be like "this is obviously why the capitalists want it" but I think it's deeper than that. Certainly the tech sector venture capitalist class has resoundingly demonstrated they don't have much interest in quality or robustness, they try to build the worst thing their customers will tolerate. If you don't believe me ask a software engineer how many large code bases they've worked in that they felt were of high quality. Then ask them if they think they're capable of writing robust code. Then ask if they think they're given the time to do so. I think it follows, then, that if some QM can generate something bad but still tolerable for a customer, in a fraction of the time they'd need to bully someone with a sense of pride into compromising their work (because I really do think a lot of people in this industry try to take pride in what they spend 8 hours a day doing to prevent losing their shit), they'd obviously do it!

There's a lot of holes in that argument and I don't think anyone with a lick of intellectual honesty actually thinks that the current state of these QMs is something acceptable even for these "as bad as the customer will tolerate" purposes. For one, they're incapable of fixing remotely interesting problems, be it in software, translation, or fucking design layout. And creating dependency on these things erodes your workforce's ability to solve the interesting problems because all those foundational skills your specific problem solving leans on have gone to rot, they're all replaced with reviewing some QM output. (Though there are professions which I think are at serious risk, and I don't want to diminish their plight)

So I don't think it's belief that you can cut labor costs by a tenth motivating these venture capitalists. They know interesting problems still require human intervention, they know human intervention is built on experience. You can't bet the farm on them eventually being able to solve these problems because you risk a gap where no one knows how to solve them, and no one who has a farm is currently betting it (though some might be betting that they can build a farm with it). Then why the fuck are they doing it?

The internet and (more sharply) the smartphone broke all these people's brains #

Silicon Valley's motto for YEARS was one of "disruption". Whatever that meant in practice (labor exploitation! rent seeking!), it was largely the belief that technology had advanced so much due to widespread internet access, be it to the home or in the form of smartphone, that the leading companies in many industries were resting on their laurels, convinced they were unassailable when an entity sufficiently motivated and endowed with resources could come eat their lunch. There are lots of non-SV examples of this... Perhaps most famous is Kodak literally having the digital photography technology that would replace them birthed in one of their own labs and electing to suppress it, only for it to be independently invented elsewhere and erode their entire business.

But internet access and the smartphone did this to like fucking everything. News, television, movies, advertising, books, music, fashion, home electronics, banking, taxis, political satire, translation, research, encyclopedias, games... If you look at the 5 largest companies on the planet by market cap, 4 are tech firms that fit this mold. Apple, Microsoft, Google, Amazon (third place is taken by the combined worth of all of Saudi Arabia's petroleum industry). Apple gatekeeps access to a third of all smartphone users on the planet, Google kinda eyes you up and down to make sure you aren't like, obviously criminal for the other 2/3rds, Microsoft and Apple similarly for most consumer software that doesn't run on a smartphone and a good chunk of professional software too. Amazon is trying to supplant physical retail as an industry, etc. Arguably Microsoft is the odd one out for the era because they haven't really replaced anything in the last 15 years, rather they adapted to how modern software, professional or consumer grade, is expected to work. I say this as someone who has an offline copy of Office (specifically for Excel) you will pry from my cold dead fingers.

The point being, these companies are taken as representatives of the level of financial return that can be achieved by exploiting (or creating and seeking rent on) new technology that ends up defining an era. On the flip side there are plenty of companies that failed because they couldn't adapt to changing times. I already mentioned Kodak, who probably could own half the cell phone camera sensor market if they played their cards right twice, or BlackBerry, who went from making the proto-smartphone to total irrelevance. Blockbuster famously declined to buy Netflix who have rendered them a piece of nostalgia. Borders Bookstore was leveled by Amazon, there are plenty of stories like these.

Go big or go bankrupt #

The SV venture capitalist class doesn't really think about the in-between results. There are plenty of businesses that are in between these two extremes of "most valuable companies on the planet" and "went out of business or at least lost lots of money," but they aren't really relevant to their mindset, and admittedly, 90% of companies they invest in fail, modest return on investment won't make up for that, their entire financial model depends on investing in a company that comes to run an entire market, or at least can be marketed as poised to do so when they go public so they can cash out.

These stories of failures to adapt are common Silicon Valley lore. The venture capitalists ensure they are on the mind of everyone they invest in because these people made bank off the last thing that defined an era. And they like to compare people favorably to the winners of that last era (everyone who does anything hardware wants to be the "Apple of X", and aiming for less is considered underachieving).

Here is where the narrative takes shape; if something new comes along, and it comes to be foundational technology for the next decade or two, the two options are your company fades to nothing, or you win big. No one knows the names of the in between players, they don't exist in the narrative. No one knows in advance if something is going to be world defining, so you have to treat anything gaining traction as though it will, or you risk the loser bucket and miss out on the pile of money in the winner bucket.

This isn't some purely academic analysis. One summer in college I had a very cool fellowship in the technology sector for underrepresented backgrounds. The organizers were very well connected in SV, so one of the things we got to do was tour Facebook's campus (I will NOT call them Meta) in like 2012. At the time, this included a former Sun Microsystems building. Sun is another entry in the tech sector failures lore (Among other things, they made Java, the language most Android apps are written in and staple of server side software. Their stock lost like 80% of its value and got bought by Oracle, long time ghouls of SV). And on the flipside, Facebook very correctly read that their products had to work on smartphones or they would not survive, they were acutely aware of the power rapid broad technology adoption can bring to bear on an industry.

Facebook had very deliberately left the Sun logos and branding up on a bunch of the glass doors and entryways in this building to, and I fucking quote, "remind us what can happen when we fail to innovate." And this reflects in how they viewed their products at the time; copy or buy literally anything new gaining traction. Acquire Instagram since it was better for photo sharing. Acquire WhatsApp since it's better for chat. Copy Vine (short form video), copy the entirety of Periscope (live video), copy Snapchat's stories feature, try as hard as you can to copy TikTok, etc. I cannot imagine the mindset and the actions are unrelated. And Facebook was rewarded extremely well for all this until 1: they failed to understand what made TikTok work, and 2: Apple demolished a huge chunk of their ad targeting business with privacy settings.

Put the fear of (not being) God in them #

The part of the tech sector that buys into the venture capitalist narrative is looking at these things wondering "What if dismissing these as a fad is like calling smartphones a fad? What if this shit makes literal dinosaurs out of us and puts us out of business?" Then they compare the cost imagined in their heads by The Fear (miss out on becoming richer than God, bankruptcy, get bought by Oracle, get dunked on by Eliezer Yudkowsky being tedious) to the cost of just fucking trying something in this space... And in their brains, this makes total sense! It's relatively cheap (read: relatively few human hours needed) for a company to plug this somewhere in their product and see what happens. Or at least, here's why I think it has to be cheap to try; ChatGPT started gaining a lot of attention in December of last year, Google had a tech demo (which admittedly sucked) this February. Google is slow as fuck at deploying new features, even half baked ones. Maybe they were already cooking on something like this and trends aligned, but between Google's speed and how fast hobbyists are integrating it with personal projects and the barrier to entry has to be quite low just to play around.

So the cost of trying out QM integration is relatively low, your imagined worst case cost is extremely high... Of course people who think this way are going to try it when enough people are talking about it. Some are more adventurous than others, some have way less to lose (or in fact are gambling with someone else's money that they can find something that will eventually go big), so they pick it up earlier. All these early stage SV founders are talking about this thing they have a surface level impression of and how cool it is, and all their friends are other startup founders or early employees, so the less adventurous hear a lot of people talking about it, recall The Fear that their bankrollers have instilled in them, so someone at their company has to try it out, it does one surface level neat thing... This repeats until even the relatively conservative institutions (the Googles of the world, or really any SV company with a real business model) hear about this and feel The Fear.

The logical conclusion to this school of thought, to The Fear, is when enough hype and attention is swirling around these things, you now have to act fast because if you look at the last 15 or 20 years of the tech sector, you could argue a lot of things tend towards duopolys (Android and iOS, OSX and Windows, Uber and Lyft,). OpenAI is obviously the first mover, and if you don't act fast someone else will be second. So you rush some shit out that you don't fully understand the implications of because The Fear is your primary motivator at this point, which is how you get Google's botched February demo wherein their chatbot got a very clear cut factual error, and Microsoft allegedly ignoring internal ethicists and launching a similar chatbot for Bing.

Silicon Valley is, fundamentally, afraid of what new technology might bring, so its companies rush to be the ones to bring it.