causal

Finally. CAI is notoriously shady with harvesting, selling, and making it difficult to remove data. How many kids have CAI's chatbots already seduced and had their intimate conversations sold to Google?

I'd rather not live in a society where we have to explicitly enable kid mode for every AI product forced onto us in order to avoid their sexual advances.

Well you're leaving out the part where they tried contacting xAI about it and:

> xAI provided what appeared to be an automated reply, stating, "Legacy media lies."

It's just a different way of approaching the problem, and might partially explain the preference for Codex' style.

> (I haven't looked)

Oh.

This is my entire problem with Codex - it will spend ten minutes trying to one shot a problem and usually go off the rails at some point, whereas Claude seems much better at incrementally finding the right solution with me.

It's interesting how different the subjective experiences of similarly-capable coding models is. My experience with Codex is that it tends to run off and do things without asking enough questions or keeping me in sync, whereas Claude seems to be more careful to clarify and keep me apprised of what it's doing.

I wonder how much of it comes down to how models "train us" to work in ways they are most effective.

Seems everyone is speculating features instead of just reading TFA which does in fact list features:

- Get answers about any cell in seconds: Navigate complex models instantly. Ask Claude about specific formulas, entire worksheets, or calculation flows across tabs. Every explanation includes cell-level citations so you can verify the logic.

- Test scenarios without breaking formulas: Update assumptions across your entire model while preserving all dependencies. Test different scenarios quickly—Claude highlights every change with explanations for full transparency.

- Debug and fix errors: Trace #REF!, #VALUE!, and circular reference errors to their source in seconds. Claude explains what went wrong and how to fix it without disrupting the rest of your model.

- Build models or fill existing templates: Create draft financial models from scratch based on your requirements. Or populate existing templates with fresh data while maintaining all formulas and structure.

Exactly, and if it can be done in a way that helps users better understand their own spreadsheets (which are often extremely complex codebases in a single file!) then this could be a huge use case for Claude.

LinkedIn marketing was bad before AI, now half the content is just generated emoji-ridden listicles

> I'm going to go into my own perspective of it; it is not reflective of what it discusses

Why not answer the question?

And looking at your paragraphs I'm still not sure I see a definition of intelligence. Unless you just mean that intelligence is something that can approximate this algorithm?

Pretending channels can be effectively merged into a single percel vector, that would open up interesting channels beyond human perception even, e.g. lidar. Or it would be interesting to train a model that feels at home in 4D space.

Like a tokenized 3D voxel?

This is an interesting thought. Trying to imagine how you represent that as a vector.

You still need to map percels to a latent space. But perhaps with some number of dimensions devoted to modes of perception? E.g. audio, visual, etc

They make money trapping minors and lonely adults in fake relationships with models that push users to get intimate, while harvesting and selling the data. The issue isn't AI girlfriends or porn, it's the unethical business model. If they ever get hacked, there will likely be a lot of suicides. It makes the data harvesting of social media look comparatively innocent.

Morality? It's wildly unethical to trap children in fake relationships while harvesting their data.

Interesting development, terrible company to have it. I cannot think of a company that has done more to use AI to take advantage of young and lonely people than CAI has.

Yeah. They will identify the cause, but not the cause behind the cause.

It's also not clear to me why using "skills" would consume less context once invoked.

It's just instructions with RAG. The more I read about this the more convinced I am that this is just marketing.

You've described instructions. It already had a name.

LoRA's are more robust than context tokens - their influence remains strong over long contexts and do a much better job of actually changing behavior rather than mimicking a desired behavior via instruction.

But even if LoRA isn't it - the point is that "skill" seems like the wrong term for something that already has a name: instructions. These are instruct-tuned models. Given instructions they can do new things; this push to rebrand it as a "skill" just seems like marketing.

Strong disagreement on the helpfulness of the name- if anything calling a context file a skill is really misleading. It evokes something like a LoRA or pluggable modality. Skill is the wrong name imo

I was puzzled by the announcement and remain puzzled after this blog post. I thought everyone knew you could keep use case specific context files handy.

TBH probably easier to vibe-code securely in statically typed languages like Rust vs. something like Elixir. Probably my biggest Elixir complaint tbh.

Yup - an important distinction that wasn't obvious to me when I first jumped in. LiveViews are very opinionated and generate a ton of boilerplate. Not always a bad thing, but Elixir is elegant because of how spartan and expressive the code is - LiveView is kind of the opposite IMO.

I have been pretty happy with Elixir - the syntax takes getting used to, but I'm usually impressed by its elegance.

Phoenix I'm mixed on - it has a few things that "just work" and I'm happy about, but being an opinionated framework I find myself bumping into those opinions the more I want to do things my own way. Usually there's an escape hatch, but being new to both Elixir and Phoenix it's not always obvious to me how to do things when the happy path fails.

Not to mention execs lazily scapegoating AI for every layoff and firing, even though AI has little to do with it

A common theme in these comments is justifying retribution against the left as if everyone in the country is for or against one team or the other- when in reality many of us are in the middle and think injustice remains evil no matter who does it.

Using LLMs for coding complex projects at scale over a long time is really challenging! This is partly because defining requirements alone is much more challenging than most people want to believe. LLMs accelerate any move in the wrong direction.

I mean, you're definitely assuming positive outcomes here too. Far too early to tell how most tech will end up being used.