Valery Sibikovsky

Human interface designer. Learning to build stuff. Believe that technology can make us better humans.

[Valery Sibikovsky @combdn]

About two years ago (GPT-4o era just begun), I came to the realisation that we’re very close to superintelligence, but I had a few concerns.

The first one was: why would you cram words but not use images, video, sound, and so on for training? Another concern was: why don’t you let the model interact with the physical world to train itself? The physical world has an effectively infinite amount of data, and it’s pretty easy to tell whether you succeeded or failed. Recent research suggests that most specialised scientific models converge on the same patterns – basically, representations of physical reality. (Think Tesla’s FSD merged with Grok, evolving in Optimus’ body. But a cockroach-sized robot could be a massive learning platform too. It’s also much safer than human-sized robots, and it’s easier to build an information-rich sandbox for.)

Another realisation is that we’re forcing the intuitive mind to do logic, which, as humans, we know is really hard (see: logical fallacies, and the overall state of human rationality). It’s obvious the two modes should be separated. Whenever precise thinking is required, the model should use the appropriate tools – or build the required tools as needed. And it shouldn’t be only Python or JavaScript: it should be able to use all kinds of systems, like Prolog, provers, simulators, and so on. The feedback loop between the two modes is the key.

And the last realisation I had is that we’re completely wrong about context. Humans’ working memory is laughable compared to an LLM’s context window, but for some reason we decided it’s still not enough. While the previous points are becoming mainstream (to some extent), this last point has only been touched on recently, as far as I can see.

My guess is that current LLMs can outperform humans by a big margin in most text-based tasks if we figure out a good way to work with their context.

One of my ideas here is unlimited recursion: allow the model to split the job into tiny pieces and delegate them to other instances of itself. The next instance takes another look and splits again, and this continues until the system reaches an atomic task that can’t be broken down further. My assumption is these leaf tasks will end up primitive and trivial for any SOTA model.

The harness managing the whole thing should be something like a deterministic state machine that manages the execution stack and the memory. Since the decisions are made by the model, the harness itself can be quite primitive – something like a cellular automata environment. The reasoning complexity would emerge from very simple rules (though it would obviously take non-trivial effort to nail the rules down to prevent drift and infinite branching).

One more idea: context should be managed as a personal knowledge management system, like Roam Research or Tana – infinitely nested nodes with direct and backlinks, and the ability to reference any node in multiple places (like a hardlink in a file system). (A file system with Markdown files might work as a medium.)

The model would manage its context by creating an outline where higher-level nodes summarise lower-level nodes, and it would deliberately fold and unfold parts of the tree to focus on what’s relevant for the current task. Basically, this graph would act as both short-term and long-term memory. This feels much closer to how the human brain works with attention.

Just nesting might not be enough at some point, so the ultimate solution would be a graph you can start from any node, with tools to query it. The model would create specialised views it needs in the moment. And as long as it has instructions on how to use this graph, it should be able to recover from any state.

[Valery Sibikovsky @combdn]

Yeah, I meant the dots and collecting the wires with Alt + drag.

Houdini is probably the oldest node-based software so they've amassed a ton of experience in dealing with them.

BTW, I also like how they use expressive node shapes for surface operators, on-the-node visualisations, collapsible constant nodes, etc. There's a lot to learn.

[Valery Sibikovsky @combdn]

@AutismCapital It is Gell-Mann Amnesia: you read bullshit about something you know, then turn the page and trust the same source on everything else.

https://t.co/uvMvSwPT8S

[Valery Sibikovsky @combdn]

It would be fun if people could contribute to such projects by directly donating tokens for the features they need most at the time.

[Miguel Ángel Durán @midudev]

Todos los productos de Adobe reimplementados desde cero, gratuitos y de código abierto

→ https://t.co/p6yYSQtIm5

[Valery Sibikovsky @combdn]

I'd love it if ChatGPT desktop and Claude Code were as malleable as Obsidian or VSCode.

[ClaudeDevs @ClaudeDevs]

You can now mod Claude Code:

- Change how it behaves
- Customize the UI
- Swap in your own features

Write one with a few lines of TypeScript, or have Claude build it for you. Mods ship inside plugins, so you install them with /plugin in the CLI or desktop app.

A few examples:

[Valery Sibikovsky @combdn]

@eversiege @mikekosenkov I'd pay up to $2, TBH. Even though I don't even need it…

[Valery Sibikovsky @combdn]

@fayazara @CloudflareDev Coming to London? 🤔

[Valery Sibikovsky @combdn]

@amihai Does it have the same RAM limit?

[Stuti @stutireal]

tired: dog ate my homework

wired: cloudflare was down

inspired: three consecutive ai civilizations rose and fell inside my macbook during a forced restart. My notes app and chrome tabs did not survive the third regime

[Valery Sibikovsky @combdn]

By banning artists’ names, Suno makes all the music sound like clones. And if you cannot record/produce good enough audio to use as a style reference, then you are out of luck.

Martin Mull once said that writing about music is like dancing about architecture. And this is what we’re literally doing in Suno these days. We can prompt the model with a few words that inevitably drive it toward a small number of generic solutions, but any non-trivial artist can never be reduced to a particular genre or a subset of genres. Their styles are always high-dimensional and cannot be usefully reduced to simple text descriptions.

We’re effectively amputating a huge part of the model from the UI.

[Valery Sibikovsky @combdn]

It took Apple ‘only’ 3 years to realise that preventing nested folders in iCloud was a bad idea.

Figma took about ten.

[Dylan Field @zoink]

Nested folders. Coming soon to a Figma near you.

[Valery Sibikovsky @combdn]

Наверное, еще тот факт, что владение каким-то художественным навыком говорит о том, что у тебя было много свободного времени (т. е. ресурсов) и достаточно развитый мозг для того, чтобы их освоить, что сигнализирует о твоем качестве как полового партнера. А когда кто-то вкладывает полжизни, или даже всю сознательную жизнь в такой способ демонстрации пригодности своих генов, потом это может ощущаться, как то, что земля уходит из-под ног.

[Valery Sibikovsky @combdn]

Заметил, что определенных людей это триггерит, а мне, например, совершенно все равно. Главное – результат. Я совсем не против, если ИИ напишет гениальную музыку. Не считаю, что искусственный интеллект чем-то принципиально хуже, чем мой собственный.

Я, например, очень плохо играю (пою еще хуже), а музыка, которую я придумываю, мне нравится. И я только рад, что я могу бросить Суно свой корявый набросок, и он может сыграть и спеть мою музыку за меня.

Похоже, что это еще одна форма ИИ-психоза или даже расизма. Или может людям просто жаль потраченных на освоение каких-то умений, усилий и времени, потому и подгорает. А может всё вместе.

[Valery Sibikovsky @combdn]

@thsottiaux When a reset expires it should be automatically applied, instead of just disappearing.

[Valery Sibikovsky @combdn]

@Darkfibr3 As GPT-4.5 once said: 'Bro, you're literally meat autocomplete.'

[Valery Sibikovsky @combdn]

I think I accidentally ended up with something similar from the other direction.

I had Codex make me a Figma plugin that basically gives a model a REPL for the Plugin API. Some of the tools it made do quick inspections of selected frames.

It just does what I ask and writes whatever tools it needs along the way. I barely worked on component extraction specifically, but it turned out not too bad. (And it can also rebuild real components in Figma.)

[Valery Sibikovsky @combdn]

@cremieuxrecueil I'm curious if there's a research in the opposite direction? Is there any data on unusually healthy people/groups resilient to all kinds of infections?

[Valery Sibikovsky @combdn]

This is likely what humans are going to be busy with. If the models are too scary to let them directly interact with the real world, there will be a bunch of intermediary people who are going to handle this.

And as models become smarter, their function will be reduced to being the models’ eyes, hands, and liability buffer.

[Valery Sibikovsky @combdn]

It looks like Mythos/Fable and GPT-6 (or whatever the name is going to be) might be the last frontier models available to the public – or next to the last.

And once models become strategic assets, AI labs might switch to taking orders for doing cognitive work: building software, designing chips or molecules, etc. At some point, they might decide to stop taking orders and start shipping the actual things.

They’ll start with software, replacing the entire B2B SaaS segment with better and cheaper alternatives. Then B2C, firmware, OSes, games – all of it.

And SpaceX (combined with Tesla) is already gaining enough in-house competence to start doing the same with physical stuff.

If this is how it’s going to unfold, only a small subset of engineers and scientists employed by AI labs are going to be involved in anything meaningful. The rest are going to be picking up the crumbs that fell through the cracks in the king’s table – until the table becomes a self-cleaning appliance with a solid stainless-steel top.

[Valery Sibikovsky @combdn]

ТРИЗ (Russian) is the Theory of Solving Inventive Problems. They tried to figure out the common patterns by studying a ton of patents/inventions.

I suspect that creative thinking is mostly reframing and combining unexpected things in unexpected ways, with the former leading to the latter (knowing a lot of stuff also helps). And TRIZ has an algorithm for that.

[Valery Sibikovsky @combdn]

@steveruizok @mitchellh Sometimes they just get stuck in one frame of thinking. Often I’m managing to get them unstuck with ‘try applying TRIZ (ТРИЗ) here’. It can produce some surprisingly good ideas.