I can’t imagine AI has anything to do with making roadmap-scribe PMs obsolete. They already were obsolete, but now there’s this idea that PMs should be prompt engineers and vine prototypes, as well as customer researchers and executive blame sponges. (And execs still probably want OKRs & roadmaps.)
Posts by David Demaree
Before I left/got shoved out of the Big Tech PM game, one of the big points of friction between me and stakeholders was execs’ insistence on having a 12 month roadmap — with *committed* OKRs — that would be stale in less than a quarter. I made the mistake of saying that was a silly waste of time.
More and more I realize that in today’s world you can only survive if you’re the type of person who loves having a spotlight shine on them and can make the most of it.
If for any reason you want to just live a quiet, private life where you don’t have to sell or hype yourself to others, TFB
Man with large head posing with his tween daughter and French bulldog
Render of a custom Funko Pop figurine styled to look like the man in the first photo
Sometimes I’m in a photo and I’m like
My god
My head is HUGE.
Does this explain why I’ve always felt so… different? All these years, have I always been…
a FunkoPOP?
Sutskever would later sit down with New York Times reporter Cade Metz for his book Genius Makers, which recounts a narrative history of AI development, and say without a hint of satire, "I think that it's fairly likely that it will not take too long of a time for the entire surface of the Earth to become covered with data centers and power stations." There would be "a tsunami of computing. almost likea natural phenomenon." AGI-and thus the data centers needed to support them-would be "too useful not to exist”
I see a lot of techies perplexed about the venom the rest of society has about AI. here’s just one sample of why so many people reflexively loathe AI
I don’t want to yuck the yums of anyone who’s excited to reinvent and put their own stamp on how we communicate online, but the vast majority of people — including highly technical people like me who are also very, very tired — just want a thing that works, ideally one that works well.
It’s worth noting that the open protocols that AtProto or Mastodon are trying to emulate — email, HTTP — were invented under totally different conditions. There never was an Email PBC, just academics on a government-created network of networks who wanted to send each other mail or share papers.
For these reasons it’s been hard for me to engage with the “fediverse” or “atmosphere”. Though I love the idea of an open protocol for social, I don’t want to spend precious energy on having to craft one, and or invent new shit that relies on one. I just want to post about video games and CSS.
Even more tangentially, I play a bunch of video games with intense modding communities, but I myself kinda hate mods. All software is kind of a fragile magic trick, and it makes me incredibly nervous to destabilize something as complex as a AAA game just for my character to have cooler outfits.
Tangential, but I’ve been trying this year to make a go of using Linux on a Framework laptop. It’s fun to basically build a whole OS from weird little terminal commands. It’s also incredibly inconvenient, and *at best* I have a weird toy system that can’t run any standard apps.
One of the hardest things to get across to most software people with great intentions is *how strong defaults are* and how tired everyone else is of dorking around trying to get things to work, but this understanding is the thing that made Apple, for all its many flaws, such a giant.
Soon AI use will be disrupted by the cheaper and more reliable alternative that's sweeping the business world: hiring human experts to do the work instead
Ironically, a lot of the tax incentives around capex assume that companies are investing cash in factories, warehouses, bricks & mortar stores — stuff that will presumably create jobs for people. The tax code didn’t anticipate a world where you could spend capital but _also_ cut jobs.
100%, but also capital spending gets better tax treatment and has less risk of liability
which is to say it’s systemic against labor in more ways than one
It may end up good in the long run that I can have an LLM generate boilerplate code; I often have it do the first, most boring 50% of a WordPress theme while I make mockups and color/spacing scales. But if Claude disappeared tomorrow, I’d go back to ‘cp -R theme-template new-theme’ like God intended
This past week I rewrote a vibe-coded app myself, because while the AI version was like 70% OK, it didn’t look very nice and I am 1000% faster at implementing CSS than any robot.
The twist: the whole thing was a side quest. I didn’t need this app, and if Lovable didn’t exist I may not have bothered
Sometimes readers are skeptical when I tell them that retailers would rather pay for self-checkout kiosks and product lockup systems and all manner of other questionable in-store tech rather than staffing because of an ideological preference for capital spending over labor spending and, well,,,
Also:
Cookie Monster
Children of the Corn Flakes
There Will Be Blood Orange
Babe-con
Add a food to a movie title:
Chicken Wings of Desire
- “Positioning” is assumed to be the answer to all marketing questions. Which isn’t entirely wrong, but easy to overthink. Sometimes your niche or position is that you’re the one with availability in your area and it’s just obvious?
- Nobody has a good answer for how to start from zero, or how to dig out of a hole if your client base disappears, other than “don’t do that”. It’s assumed that you, the founder, are starting with momentum and connections. (This kinda makes sense — otherwise why would you start a business?)
- Lots of people claim to have systems or frameworks for “agency growth”. Some of them may even work. But they all assume you have a steady pipeline of opportunities and just have trouble keeping track of and optimizing all your money.
Some observations after eleventeen or whatever months as a “web agency owner”:
- “Agency” is assumed by a lot of folks to mean “team of randos who build stuff” and the first questions will be how many randos and how much you charge
- Most clients *and fellow owners* are defiantly non-technical
hello 👋 i have developed a new tool called blüuümpíi. blüuümpíi will convert every file on your hard drive to a .wav file without asking you, maximize your system volume, and play them on every bluetooth speaker you have ever connected to. if you offer feedback on blüuümpíi i will take legal action
What are you talking about? That dispensary has been open for months. You’re just imagining the paper in the windows.
I realized earlier that there are two big reasons why Starfield wasn’t a giant megahit on release and also seems to piss people off:
1. it’s a game by and for 40-something nerds who grew up on 1990s sci-fi
2. If you’re gonna promise “Skyrim in space” a lot of those nerds are gonna hold you to that
It’s bad enough to have to worry about AI-generated code being insecure, not scalable, poorly instrumented, etc, but you also have to check whether it actually does its job or just wrote a placeholder function and forgot it wasn’t the real thing.
It’ll have implemented a function called “sendEmailSummaryWithPDF”, and the function may even do something, but not sending emails or generating PDFs. And you’ll find this out when a user sends a bug report that their email/PDF never arrived.
One of the under-reported issues with vibe coding is the thing where the LLM says it built something — say, a feature that generates a PDF and emails it to the user — but is just lying.
This is an awful lot of words for Grammarly’s CEO to use to say “you get paid in exposure.”