garethsprice

There's a neat non-profit in Upstate NY, Holocenter, "center for the holographic arts". They have exhibits of currently working holographic artists, and you can do a workshop to make holograms: https://holocenter.org/

That's a big if. The largest technological improvements (electrification, internal combustion, antibiotics, plumbing) mostly arrived before the post-1950 population explosion, and productivity growth slowed right as population growth peaked in the late 60s. What drove tech innovation was research and R&D spending, much of it funded by governments as part of war efforts.

Wait, now I'm nervous...

That's fast turnaround! Thanks!

Thanks for building it! Looking forward to seeing how it - and people's setups - evolve.

1. As a co-founder, there is not much time for hobbies and that is part of the deal. Plus local inference is my hobby, at least at this point in the game. The free time I do get I prefer to spend offline and outside with my partner. I have asked Claude for relationship advice (which it is good at) but no workflows around this (and I think that would not go down well).

2. Early days, the tooling is not quite there yet. I think this is the future of engineering, the days of hand-crafting code - and probably reading it - are gone forever. It generates a LOT of tokens, but that is OK when they are cheap enough as long as the result is consistent. Speed is not so important which is great for local inference - can leave it running overnight and come back to a bunch of completed work even if it took hours.

I manually tag a ticket for automation in Linear. Experimented with having a separate skill that looks for automatable tickets and auto-tagging them, worked well, but having enough work for it to do is not the bottleneck compared to 1) having enough triaged, well-written work ticketed out, 2) reviewing its output. There's articles out there about arranging work into tiers based on complexity/risk, with the lower-end being fully automated to free up time for humans to focus on the high-end. Not had time to work on this much but that's probably where I will go.

A second human has to approve every PR in our repos (good practice and also a requirement for SOC2, which will impede any fully-automated pipeline adoption in regulated industries). I also do a first-read of the draft PR to make sure it is good quality, as I am ultimately responsible for the work my agent creates and to shift the responsibility of first-look (for any AI output) to another person is disrespectful to their time and inefficient.

Hence "glass-factory" as the "Dark Factory" pattern is a black box with no human intervention, but I think that is not something suitable for how teams build software products for real users presently, at least in the near future.

3. Forcing something on people is a good way to have them reject it. I hire professionals to achieve a goal and they have agency in how to achieve it. I encourage everyone to share their work internally (demos, repos, etc) and things that are truly beneficial to people's workflows gain traction quickly. As a result we have many experiments and a culture of rapid innovation that encourages people to try new things.

Negatively, this creates a proliferation of wheel re-inventing that eventually benefits from consolidation (eg. do we really need 3 people's personal apps monitoring ETL pipelines). But we are also in this emerging age of personal software where it seems to make most sense to consolidate at the API or documentation layer and let people - engineers and even business-side now - continue to vibe-code their own workflow-specific apps.

The precedent has been to build generalized software for an audience as it was expensive and time consuming to build, but when software is cheap and fast to build it may make more sense for the user-facing application layer to be made up of many small ephemeral applications that are only useful for one individual, that change rapidly as their needs change.

tl;dr: Not forcing anyone to use it as it's still experimental, yes personal experimentation is hopefully the basis of successful ideas that will be adopted by the team when they provide real benefits to their work.

Thanks! No repo yet, it's very specific to my use case. I'll update here if I ever publish something, but here's a summary that should be enough to vibe-code something similar.

The audio UI (AUI) is a small client that consumes the event stream the factory puts out (same one the TUI client uses), it runs on my desktop machine which has speakers connected. It works fine on good laptop speakers, but it benefits from a good subwoofer as it works best with sounds that are on the edge of perception.

A low rumble at a base of 55hz plays through the subwoofer to indicate that the factory is active and idle, low enough to fade into the background but noticeable in its absence (ie. the factory isn't running). A heartbeat on the feed announces the tok/s volume every few seconds which adds harmonics on top of the 55hz, 0.5hz at a time with a 15hz cap on the swell so it accelerates gracefully. Loudness is increased with parallel tasks, creating an X and Y axis (tasks x token volume). This creates a non-intrusive but interpretable signal for factory activity that becomes intuitive and subconscious. The rumble sounds like a distant car idling, warm and comforting so it doesn't become fatiguing over a whole day (research on low frequency brown noise shows it is good for reducing anxiety and improving sleep).

Events trigger sounds. Sounds for previously unseen events are generated with an LLM and numpy (eg. "make a video-game style sound effect suitable for an {announcement that a PR has merged successfully}") to output a waveform and save it to a lookup table so they are consistent over time. Each sound has a rate limit and a cooldown period so it does not become annoying. This table can be tweaked manually if generated sounds are not good. A mixer module controls the levels so that important sounds are louder and less important or frequent ones blend in the background more.

The human brain is primed to interpret environmental sounds so you quickly learn what each sound means and it becomes almost subconscious to follow what is going on at a level that is deeper and faster than interpreting text output. New sounds definitely trigger alertness and I can go see what just happened. Just playing sound effects gets annoying quickly so it becomes almost a composition exercise to create a long-form soundscape that is novel enough to communicate but not fatiguing (I don't have any musical abilities).

Some events can trigger text-to-speech (TTS) via whisper, eg. when a draft PR is ready for review it plays the sound effect (a chime) and announces the ticket number and title is ready. A blocked step is a dull thud and a 5-10 word description of what's blocking it. There are "quiet hours" where it does not speak, and when it wakes up it gives me a spoken digest of everything it did overnight and what needs my attention.

I connected Piper so I can talk back to it ("factory, what just happened?") but mainly to see if I could. I haven't built a habit around it, I got into this job because I prefer typing anyway.

A lot of this is based on ideas from ambient music (Brian Eno etc), the MIT Media Lab's work on ambientROOM (1997): https://tangible.media.mit.edu/project/ambientroom/ and a somewhat unfashionable belief in Steven Levy's hacker edict that "You can create art and beauty on a computer". LLMs have made it possible for me to build out ideas that have been floating around in my head for a long time. This is an exciting time to experiment and build new ways of doing things. Hope this inspires you to build something similar.

Nice idea. https://mysetup.ai/u/gareth - tried to include useful hints for others (screenshots, token/sec results, monthly costs, etc).

Interesting to see how many people are spinning up custom workflows and factory-type patterns that run alongside AI coding tools (myself included).

Would love to see this collated into a regular survey to pick out trends. And an RSS/Atom feed or API so I can have an agent watch it :D

The added speed of AI tools means they're now able to predict 18 of the next 10 recessions.

Zack London's recent cyberpunk short "The Patchwright" (21 minutes)[1] feels like a glimpse into the AI feature film future. I like that rather than trying to replicate a traditional film it leans into the maximalist AI aesthetic.

1: https://www.youtube.com/watch?v=-Rzl7nUdEs4

I disagree that ordinary people marveled at electric motors and saw their potential. Electric motors were seen as a toy for decades following their invention in 1834, and required industry to completely re-design its approach before they were used in more than 5% of factories. Even then the big shift to electric from steam in industry - almost a century later - was more due to economic necessity than better technology: https://www.bbc.com/news/business-40673694

I think the same cycle is playing out with LLMs - the general public is somewhere between indifferent and actively hostile to them, as they have been with every new technology. Reconfiguring businesses and society to use them to their full potential - and benefits to society - will take years, if not decades.

There were cryptic billboards in the last 2 tech booms too (dotcom, web 2.0) - the one that comes to mind is the Google ones that were a puzzle that led to a recruiting funnel (2004: https://www.npr.org/2004/09/14/3916173/google-entices-job-se...)

Trash AI article with a useless map. It looks to be a screenshot from an interactive version in the linked Newsweek article[^1] that was mangled by the slop grinder.

The source map shows current electricity costs rather than projected ones. This makes more sense given NY just implemented a moratorium on new data centers (because who uses the Internet in New York anyway /s) but it was dark blue on there.

[1]: https://www.newsweek.com/map-shows-electricity-costs-in-ever...

n8n pipeline that scrapes industry sources daily, uses an LLM to score each story according to how relevant it is to my particular focus of interest, and sends me a weekly digest on Friday mornings. Could probably do something similar with OpenClaw or scheduled Claude tasks or a bunch of other ways pretty easily.

General old-school way of not being overwhelmed by tech news: just picking stuff up when I need it to solve a real problem. News in general: is it genuinely improving your life to spend time on consuming it? If you cannot name concrete life improvements, cut it out entirely.

2014-15 was when the news feed switched from static, human-written code to machine learning models optimized to maximize ad revenue[1], the models discovering "user attention time" was a far better metric to optimize for than clicks, likes or shares[2]. And what people spent the most "attention time" on was content that triggered raw moral outrage, fear and tribal hostility[3].

This switch does appear to have put off a lot of smart, thinking people as the newsfeed degenerated into a torrent of dopamine-spiking diarrhea. But those people don't tend to be the ones that click on ads.

[1] https://time.com/collections/the-cloud/3950525/facebook-news... [2] https://www.theguardian.com/technology/2015/jun/15/facebook-... [3] https://www.supersummary.com/the-chaos-machine/summary/ - the whole "Chaos Machine" book is excellent

Same - at the end of the day I'd leave my last turn running with something like "I am going to bed so keep going until you are done" and be surprised that in the morning it'd kept going.

This latest crop of frontier models (Fable, Sol) are very good at broad exploratory directions like "audit for bugs" or "find places in this codebase that I custom-implemented functionality that has since become available in the libraries I am using" in ways that prior generations would fail or not return as high quality results.

I did a recent head-to-head comparison between Qwen 3.6 35B running locally and Fable on the above and while Qwen did an admirable job finding a good handful of things, Fable returned many more, and more serious, issues. This included many more issues that crossed system boundaries in a multi-service repo.

In terms of one-shot "follow these detailed instructions" I think we are at a point where frontier models are overkill, especially at the price. For broad fan-out explorations, I think we are just getting started.

Many harnesses include a current date and time in their system prompt, and if there is a way for the model to call for an updated time (either a dedicated time tool or calling the OS' `date` tool) they can track time they spent doing something. If not told up-front, they can try to infer it from timestamps in their logs. Sort of like a human - if you ask them to time something and give them a stopwatch, they do it. If you ask them post-facto they'll estimate it.

This "spend at least 8 hours" trick is a new one to me, though.

https://www.pckeyboard.com/page/category/Buttons <- these also fit the IBM ones (still using the Model M I pulled out of a telco dumpster in the rain when I was a teenager, it's a bit Ship of Theseus at this point with bolt mods and USB but still the same to me)

Good essay. In my own readings on art and artists, one overlap with substance abuse is that artists tend to be highly sensitive people - the type of people who are drawn to both creative pursuits and also to coping mechanisms like drinking to numb the over-stimulation.

So it is not so much that writing leads to alcoholism, so much as being the kind of person who is deeply sensitive to the world around them makes both of those activities appealing.

As heavily drinking becomes less socially acceptable, and people learn more about mental health and healthier coping mechanisms, I wonder if there will be less famously intoxicated public figures - and if art will be better or worse for it.

Using OpenCode + OhMyOpenCode + Qwen 3.6 35B-A3B Q_4_KM on an Ada 4000 (20GB VRAM) at 55 tok/sec for generation (slower than it sounds as OpenCode has a bunch of context it adds). Meaning to check out pi when I get a minute as I hear that one mentioned a lot lately.

I am using Opus to generate plans that the local agent then follows, then validated by Opus. So I'm not at 100% local but these models are increasingly part of my production workflow. Probably not worth doing - yet - unless you are a hobbyist who likes spending time and money tinkering.

This setup is certainly not as "good" as Opus or other frontier models but they are "good enough" for an increasing number of rote tasks. You don't need to drive a Rolls Royce to the supermarket, when a used Corolla gets you there just fine.

It also enables new workflows that would be cost-prohibitive with frontier LLMs (especially as token costs rise) - eg. overnight I use the Chrome devtools MCP and have the above setup fuzz-test as a user for a number of hours and see if it can break things. Even got it working with multi-modal so it can check screenshots, which blows my mind (and not my wallet, as Claude+screenshots burns $$$).

The "12-18 months behind frontier" sounds about right, it's about where I was with gpt-4o and basic harnesses back then. In another 12-18 months my bet is we have Opus-level models that can be run locally for <$5k... but the frontier models will be even further forward (unless governments have blocked them). Fun times.

One part of ADHD is hyperfocus. ADHD is a dopamine deficiency issue, so when something highly stimulating (like a good book, or writing code) causes a release of dopamine the brain locks onto it and won't let go.

https://www.fidonet.org/inet92_Randy_Bush.txt - contemporary account that is still a pretty good overview

https://en.wikipedia.org/wiki/BBS:_The_Documentary has an episode entirely on Fidonet (creative commons, freely available online)

Memories of being amazed to see messages from all around the world arriving to my hometown's local BBS, merely days after they were posted.

You are, I see, a member of the "reality-based community" as Karl Rove presciently put it back in 2004. You live in a mental space where you critically evaluate facts and make decisions aimed at maximizing your outcome in life.

You would do well to explore this great nation of ours and realize that a wild and terrifyingly large number of its inhabitants travel in an almost unimaginably different headspace.

The men who roll coal are not doing it in cheap beater cars. A Ram or F-150 Super Duty is a $40-80,000 vehicle which requires $200-$5,000 in modifications so they can pull up next to a Prius ($28-36,000) and belch out thick black lib-ownin' ball-crushin' smoke, damaging their own engine in the process. You may ask them why, to which they'd spit a wad of brown tobacco in your face and scream in primal rage into the air: because f*k you, that's why.

It's hard to calculate the number of people who don't subscribe at all, but you can calculate the number of recovered subscriptions from a retention process.

FWIW, as a subscriber to both, I have more often found myself manually renewing lapsed subscriptions than going through painful cancellation processes to get out of them. I get the Economist through DiscountMags.com where it is often available with a discount.

The book "The Paradox of Choice" led me to using Wirecutter and Consumer Reports in recent years - if it's not something I care that much about, I will just buy whatever their top recommendation is without second thought. The cost of the subscriptions pays for a human to wade through all the junk and select something that is at least adequate for the job, and is cheap compared to the effective hourly rate of having to search through mountains of junk.

For more niche stuff, Reddit groups like BIFL and seeing what niche hobby groups coalesce around is helpful for me. Which I suppose is the same reversion to word-of-mouth expertise when, as you said, brand and price are degraded signals.

Watch the ageism there, older devs can be lazy and ignorant of security too! (And are responsible for building a dev process that catches such things in review - which points to larger systemic issues over there)

I will agree that anyone that works at Meta is likely not somebody who values privacy very much, though.

Hah thanks. I also feel this way rage-spotting AI-generated text everywhere these days. I have been looking into the speech patterns of LLMs lately ("unslopping" my own AI outputs) so had a bunch of the most horrible cliches memorized. Unfortunately this is a useless talent as a "reverse mechanical turk" at approx $25,000 for 1M output tokens (at 0.3 tok/sec) is not competitive in today's market.

Yes: https://arxiv.org/abs/2409.01754

You're absolutely right to call this out — and honestly? I want to sit with that for a moment. Here's the thing: this isn't really about AI writing. It's not even about coding agents. It's about something much deeper. What's genuinely worth knowing: while I generally agree, many people may not. I think there's a really interesting conversation to be had here. Thanks for naming this. It needed to be named.

(/s - Blargh, writing like that that by hand is exhausting)

From the article:

> The brains are already almost devoid of the coordinated neural firing necessary even for minimal consciousness, says Brendan Parent, a bioethicist at New York University Langone Health and one of six ethicists on Bexorg’s advisory board. But the company also forestalls any electrical activity with the anesthetic propofol, among other measures.