technoplato (@technoplato) profile picture
Joined April 2011
technoplato retweeted
AnySpeechRecognizer implements Apple's new SpeechAnalyzer API with a choice of engine. Change `import Speech` to `import AnySpeechRecognizer`, and run NVIDIA's Parakeet or Alibaba's Qwen3-ASR locally on Core ML. It also adds speaker diarization, so you know who said what.
3
3
1
13
2,399
technoplato retweeted
i built a skill that makes any app ready for the new iphone duo most apps will just be stretched to fit the bigger screen this skill actually redesigns them for both displays here's what it does: → designs your app for both folded and unfolded screens → creates stunning UI concepts from your references → adapts existing apps built with swiftui, react native, flutter, etc. → handles the transition between both screens → generates apple-style product mockups of your app works with claude code and codex and i’m giving it away completely for free retweet + comment "duo" and i'll send you the skill (follow so i can dm you)
129
77
8
403
35,316
technoplato retweeted
If you have Anthropic, Grok, or other non-OpenAI tokens to burn, I could use some help! I'm bumping up against a TypeScript performance challenge with Foldkit's Block composition API, currently in preview. We're asking TypeScript to do a lot of inference, and some editor operations take several seconds after an edit. We've found changes that improve performance (including API changes), but I'd love to retain the proposed shape if possible. I've had Astra working on this at Max effort for the past day, and we still don't have a complete fix. Perhaps Fable 5.1 or another frontier model will be up to the task. Or perhaps tsrs will rescue me? This issue links to a reproducer, measurements, and the approaches my agents have already tried. Thank you! github.com/foldkit/foldkit/i…
9
5
1
34
2,717
technoplato retweeted
everything that can be will be converted into `state = memo { f(log) }` all other state management is just too complex at the limit
Something new we've been cooking on: 🍳 Overeasy: A durable filesystem overlay, backed by immutable logs on S3 Overeasy makes your FS: - forkable - resumable across machines (demo 👇) - revertible to any prior state
43
41
22
910
230,817
technoplato retweeted
Ughhhh I spend so much time booking flights and haircuts and ordering pizza I hate it so much agggghhhhh if only a superintelligent entity could do it for me :(((((
4
2
48
1,198
technoplato retweeted
The Death of Software.
1,096
17,071
1,243
144,661
8,225,702
technoplato retweeted
Introducing Griffin, the first model to pass the video Turing test. 48% of people who talked to it live thought it was a real human. Previous systems have had a pass rate <3%. It is #1 on NVIDIA's benchmark for full-duplex AI video. It’s the first Human Interaction Model (HIM).
Readers added context they thought people might want to know
The 48% figure and "video Turing test" claim are from Tavus's own study of 54 one-minute calls, not independently verified or using a standard protocol. Griffin-Lite leads NVIDIA's VideoFDB benchmark on their public leaderboard. cellcog.ai/blog/tavus-gri… research.nvidia.com/labs/amri/proj… tech-ish.com/2026/10/02/tav…
3,440
4,278
5,644
39,182
21,218,952
technoplato retweeted
Just shipped this update to my open source Apple HIG Agent Skills project. This update brings the /duo command, which takes your iOS app and redesigns it for iPhone Duo, using Apple's guidelines and recommended process. github.com/justinwetch/HIGAg…
20
25
3
442
202,292
technoplato retweeted
Highly recommend routinely brain-dumping to your agent of choice via dictation, asking it to reflect your thinking back to you in extended form, then using ElevenReader to listen to it while doing chores, taking a walk, etc. Been hugely helpful for me the last year. Especially helpful if your agent has long-term context on you and the project.
3
2
32
1,016
technoplato retweeted
Continúa la saga ahora con el diagrama de un modelo de difusión, a manos de Opus 5.5
I'm sooooooooo in love with this ❤️❤️❤️❤️
33
75
5
1,658
71,509
technoplato retweeted
Model your agent logic as a state machine. Give Jev the state + goal; the enabled events it can trigger are Choices. That's how you use Jev with state machines. It works extremely well.
36
34
3
742
32,731
The world changed forever in 2020. Millions of people realized they had been sold a lie. The dollar wasn’t safe, the government wasn’t coming to save them, Wall Street was out to profit off of them, and the media was lying to them. This caused people to leave their jobs, move to new cities, and question everything they were being told. This generation is never going back to believing the lies. They realize they have to take control of their life, including working at jobs with financial upside, making their own decisions in their investment portfolio, and never counting on government programs like Social Security. Each person is pursuing true independence. The “experts” aren’t any smarter than them. Anyone can learn anything with the internet and a little curiosity. There is no more handing over important decisions related to finance or healthcare to others. Pure individual sovereignty. I would never bet against this group of people. There is nothing more powerful than an informed individual who is motivated to build a better life for themselves and their family. It is these exact people that I call the misfits. They don’t fit in to the normal system. That doesn’t bother them though. They are proud to be different. Proud to be resilient to the mainstream bullshit. They are 100% hardcore themselves. They may be different, but they are right. But being different and right is where all the value lies in life. If you are one of these misfits, come join a group of misfits at @cfosilvia. We have been waiting for you.
85
48
5
663
74,292
technoplato retweeted
How “solved” does programming need to be before we get fast & small & working software?
79
65
9
1,716
48,763
technoplato retweeted
Replying to @mattpocockuk
If your software can watch itself you don’t need to Architectural avalanaches need to be moved or melted before any of this AI can truly benefit us. I’m working on prototyping that here: teapawt.net/technoplato/status/209… PLATONICALLY IDEAL SPECIFIED SOFTWARE PREDICTABLE OBSERVABLE UNIVERSAL REPLAYABLE Introducing: Piss Pour Software, Industries
I'm putting together a prototype of an architecture where you write your business logic once and can render it on whatever funky Javascript framework you choose! Built with @foldkit and @instant_db by the wonderfully talented @devinjameson and @stopachka (and team; respectively), I look forward to building out an idealistic architecture for all of the future applications I build that gets rid of the drudgery I've experienced developing frontend applications. I also envision something like this can begin to dissolve the "client/server" barrier that I feel like is just a relic from the past when space/memory were limitations we had to deal with. You will be, of course, still able to run this on some server, but the point is that if the client knows all of the business logic, it can behave offline and defer to the server super heavy jobs that aren't well suited for client and edge devices. Of course, this is just a demo of some counts and navigation state synchronizing, but the fundamental bones here are going to really start to yield a good return on investment. Your data is also going to sync in real time via InstantDB (or whatever other sync engine you choose) Anyway, enjoy and thanks for watching and I'd love to work with you if this kind of thing is interesting. #localfirst #softwareshouldwork #syncengines #distributedarchitecture #tea #theelmarchitecture #dohashtagsevendoanything? #dontcareimexcitedtoshare Also huge shoutout to @grok and @cursor as they've been immensely helpful in building this out where @claudeai and @ChatGPT struggled or just took forever to get things done. youtu.be/bGW7YlSGjgE
1
40
technoplato retweeted
Having worked at large corporations and having repeatedly failed to bear their strikingly Kafkaesque torments, I have come to regard it as self-evident that "productivity" is not even a quaternary concern of such constructions. These are vastly involuted organisms, ruled by an innumerable and unfathomable set of emergent properties, populated by men, yes, but wholly insensible to the infinitesimal whims and futile provocations of their human substrate. Despite the occasional mockery thereof, I find the near-tautological maxim “The Purpose of a System is What it Does” to be a useful pocket-sized reminder to disregard the sloganeering and self-conceptions of these inherently inscrutable orders and their well-inculcated acolytes; to behold their manifestos, mission statements, alliterative enumerations of corporate values, and even the very Words on the Tin with droll indifference, as the self-delusion, newspeak, and indoctrination they so nakedly are. I've seen things you wouldn't believe. I've seen projects developed by no fewer than 50 engineers, spread across multiple teams, plowing headlong into nothingness for years on end without anyone seeming to know for certain whether the object of their exertions was either already in production, just a proof-of-concept, or perhaps a fork of some abandoned predecessor. Now for the hope-inducing section: Many take it for granted that AI poses a threat to white-collar jobs, perhaps as a looming extinction event for this entire caste. Yet this rests on the implicit assumption that such jobs are part of something more logical, more coherent than a stultifying weave of make-work flesh drunkenly grafted onto a puke-slick, palpitating mass of sinecures whose unknowable and alien aim includes, as a byproduct, the sorting of humanity into soul-splutteringly inefficient hierarchies of asexually self-replicating scrum masters and mid-meiosis, daylight-balanced light illuminated middle managers, themselves configured primarily for the incessant heaving of damp air to and fro upon begrimed lenovo thinkpad zoom-call pockmarked stand-up screens, ouroboric and onanistic in their encircling of a charybdis of story points (which occupy a superposition of effort, complexity, and time - and the scrum people will actually bite you if you try to pin them down on this) and so on and so forth. So best of luck to AGI with all that shit. I forget my point. Do you have any blockers?
33
15
4
256
17,794
Would love to discuss, at the moment its like that😁 I’m building the first non trivial app with it and made me reconsider some ideas (to be more agent friendly) but I’m close to have something usable to share
1
1
1
91
technoplato retweeted
Introducing tcut terminal videos as code. write 10 lines of typescript → get an mp4, a gif and an svg Same script, same pixels, every time. Built on Bun + Ghostty. MIT bun add -g termcut - tcut.amanv.dev
18
25
6
274
24,432
Oh nooooo… this can’t be happening 😭 @instant_db - the database I was using for @owndotpage and many of my other projects - is shutting down. Instant has been one of the coolest startup products I’ve used, backed by an ambitious team and vision. I’ve recommended it to everyone looking for a great database. I’m genuinely sad to see it go, and I wish the team every success in whatever comes next. Migrating my users and apps will be a huge undertaking, so I’ll probably self-host Instant instead. Still, I’m curious: What other databases would you recommend, and what do they cost?
39
3
5
57
9,885
technoplato retweeted
I’m tired of being asked to review some document that is actually just ai slop. How does one avoid this!?
275
11
12
500
52,810
technoplato retweeted
codex six hours into a 15 minute task
116
430
77
11,306
600,723