Sam Cymbaluk

Co-Founder @BuildCoherence | Building Axiom to elevate logical discourse and knowledge 🚀 | Husband of @RealMetaBias

[Sam Cymbaluk @SemanticSamuel]

In Axiom, arguments are proven valid in Lean. Here, if you accept the 6 premise nodes, you must accept the conclusion.

The point is not that this is a perfect argument, it's that we can use this imperfect argument to pinpoint where our reasoning needs to be improved. Readers in Axiom flag which premises they disagree with and the author can then work on either changing their argument to be about a weaker claim (e.g. "We mustn't eat factory farmed meat"), or deriving their original premises as a conclusion from less controversial premises.

[Sam Cymbaluk @SemanticSamuel]

What's your favorite infohazard? Mine is quantum immortality.

[Sam Cymbaluk @SemanticSamuel]

AI won't replace you, not because layoffs due to AI aren't real, but because replace is the language of a zero-sum substitution. It reflects a fundamental misunderstanding of the transformation we're undergoing.

[Sam Cymbaluk @SemanticSamuel]

@DanielMiessler Yes, LLMs predict the next token but only because this is the way we trick them into learning how the world works. Cool demo!

[Sam Cymbaluk @SemanticSamuel]

@_rygo6 Yes, the exhaustive prior art research is great. That and doing frequent "spikes" to prototype different approaches before deciding on a design allows me to produce better code than I could've pre-AI.

[Sam Cymbaluk @SemanticSamuel]

@JustDeezGuy People definitely don't take the Curry-Howard correspondence seriously. You can tell because we have still have two seperate words, logic and computation. Acceptance of the correspondence would mean having a single word that semantically covers both.

[Sam Cymbaluk @SemanticSamuel]

Fitting branch name for work on formalizing language. I think Newton would adore how far humanity has come since he started us down this path.

[Sam Cymbaluk @SemanticSamuel]

@atmoio The flaw I see in your "no shortcuts" thinking is that AI isn't starting from year 0, it's starting from year 4 billion with us at the helm. We have the results of the search to learn from. Thanks for thinking in public. It's a valuable to have your long-form perspective.

[Sam Cymbaluk @SemanticSamuel]

@rob_mcrobberson I appreciate your perspective. As a first time founder, I'm learning how to pace myself and prioritize my long term health. I also have kids, so 100 hour weeks are doable as one offs, but not sustainable.

[Sam Cymbaluk @SemanticSamuel]

@ForrestPKnight Remember when 1 1 1 was the main way to code with AI?

[Sam Cymbaluk @SemanticSamuel]

There's a shocking difference in CO2 levels running a "portable" air conditioning unit in one room of our house. Modern AC installations focus purely on energy efficiency, but the tradeoff is real. For my family, fully functioning brains are worth the extra energy cost.

[Sam Cymbaluk @SemanticSamuel]

@gerardsans Good thread of agent failures. How these ground facts translate to market impact is the big question. Agents autonomously writing complex code doesn't mean all coding jobs are done for. Conversely, bad AI failures don't imply there won't be lots of coding jobs made obsolete.

[Sam Cymbaluk @SemanticSamuel]

@gerardsans Good thread of agent failures. How these ground facts translate to market impact is the big question. Agents autonomously writing complex code doesn't mean all coding jobs are done for. Conversely, bad AI failures don't imply there won't be lots of coding jobs made obsolete.

[Sam Cymbaluk @SemanticSamuel]

I'm seeing a lot of negativity about AI agents for coding. Some of the hype pullback is healthy; delusional optimism is bad. But don't lose sight of the magic!

Yes, we are all now very familiar with all the ways agents fail, but the kind of things they CAN do are incredible. Go deeper, use the failures to understand the tech better, and reinforce the patterns that work.

[Sam Cymbaluk @SemanticSamuel]

@RoseSilicon I see magic as simply technology that we don't understand very well yet. And with how fast AI is changing, we'd be wrong to think we understand it well. We should be excited to try to push the limits to make the most of it, not discouraged when we find that they exist.

[Sam Cymbaluk @SemanticSamuel]

@DaytonEllwanger I've seen stuff like this from a few devs, but for the most part, I agree. https://t.co/Ef2ukE9S4z

[Jussi @jussisaur]

i am quite close to going back to an autocomplete-only AI coding style. dead serious. i'm not sure the ostensible speed of agent-first coding is worth the brainrot, the laziness and the loss of code and architecture comprehension

[Sam Cymbaluk @SemanticSamuel]

This is probably the most magical agentic coding setup I have working so far. Tag an issue and an hour later a senior quality task or bug fix PR is ready for review. It's grounded in our in-repo docs and often finds other areas for improvement beyond the issue scope.

[Sam Cymbaluk @SemanticSamuel]

@JohnnyNel_ Yeah 4.7 has no problem with this. It identifies the soundness bug first turn, and it actually checks for lean and even tries alternative approaches when it runs into the install block.

[Sam Cymbaluk @SemanticSamuel]

I haven't noticed much of a difference between Opus 4.7 and 4.8 in claude code, but 4.8 in regular chat is dumb as a stump. I feel like I'm back in 2024. It's so bad I assume something must be broken with the CoT (I'm on high thinking mode).

[Sam Cymbaluk @SemanticSamuel]

@Westoncb Cool! LLM interfaces are nowhere close to settled and I love the idea of exposing structure visually. What do the arrows going down and right do/signify?

[Sam Cymbaluk @SemanticSamuel]

@yacineMTB If a social media platform relies on forcing slop on you to stay afloat, maybe it should go to 0.