rocho

I've been using Fastmail with my own domain for a few years and never had a problem. Web and mobile apps are great.

I've set up IMAP syncing with my main Gmail so that it also acts as a backup. Eventually I'll ditch my Gmail for good, but sadly I'm not there yet.

Gemini models have consistently been the best, by far, at writing and explaining.

Indeed, that's what I was getting at in my last paragraph. Businesses invariably determine that speed of delivery is a superior metric to optimize for, rather than learning and maintainability. In academia the situation is not so clear cut, but there are still powerful incentives to deliver faster. We'll see what happens.

As a software engineer, I found that a codebase that is LLM-assisted or entirely produced by agents gets harder and harder to understand intuitively over time. By writing code manually, you build a very precise mental model of how it works. The time it took was the price for that. Now, the time cost has been largely removed but the payoff also disappeared: it's essentially impossible to maintain the same mental model with the speed and volume of AI contributions. Reviewing code is not a substitute for writing it, and new contributions to the codebase are essentially bound to LLM usage. So now we're in a position where the codebase and its capabilities grow fast, but are not backed by the full understanding of the engineers. AI tools have become our crutch.

Math is very similar, and requires an even higher degree of mental understanding to produce advancements. The effort spent trying to prove theorems and connect theories pays off in deep intuition; removing that struggle significantly impairs our understanding.

So no, you can't "be done with that, and go do other great work" without paying the understanding tax. And over time that's just not sustainable.

A simpler analogy would be gym training: the effort expenditure and the strength gains are inextricably linked. You can't go in there and start lifting weights with a forklift while expecting strength improvements.

In the case of software development, businesses have decided that speed gains are more important than maintainability, understanding, and independence. Will the same happen in math? The future in which mathematicians become LLM assistants (as opposed to LLM-assisted) and spend their days "turning the crank" just to review what comes out is quite bleak.

Strawman. There's very little pressure, if any, to drink all the time. On the other hand, businesses are pressing employees to constantly use AI tools. The sad fact is that in many cases, they do result in faster work, but at the same time they carry so many negative effects for the individuals.

Why do you consider bonds usury?

I absolutely agree, although even that doesn't solve the root problem. The underlying LLM architecture is fundamentally insecure as it doesn't separate between instructions and pure content to read/operate on.

I wonder if it'd be possible to train an LLM with such architecture: one input for the instructions/conversation and one "data-only" input. Training would ensure that the latter isn't interpreted as instructions, although I'm not knowledgeable enough to understand if that's even theoretically possible: even if the inputs are initially separate, they eventually mix in the neural network. However, I imagine that training could be done with massive amounts of prompt injections in the "data-only" input to penalize execution of those instructions.

I find Gemini is outstanding at reasoning (all topics) and architecture (software/system design). On the other hand, Gemini CLI sucks and so I end up using Claude Code and Codex CLI for agentic work.

However, I heavily use Gemini in my daily work and I think it has its own place. Ultimately, I don't see the point of choosing the one "best" model for everything, but I'd rather use what's best for any given task.

That's not DeepSeek, it's a Qwen or Llama model distilled from DeepSeek. Not the same thing at all.

You don't need to propagate it, you just need to show the gradient of the current position alongside with the classical evaluation, to give more context to the viewers.

That's correct. I saw a paper recently that showed how LLMs performance collapses when they are trained on synthetic data.

Why would they be? Cursor took an existing editor and added some AI features on top of it. Features that are enabled by a third party API with some good prompts, something easily replicable by any editor company. Current LLMs are a commodity.

To you maybe. But if Claude or any other competitor with similar features and performance keeps a lower price, most users will migrate there.

See "AI models collapse when trained on recursively generated data"

https://www.nature.com/articles/s41586-024-07566-y

But that does not prove anything. We don't know where we are on the AI-power scale currently. "Superintelligence", whatever that means, could be 1 year or 1000 years away at our current progress, and we wouldn't know until we reach it.

But a genetic analysis to find your "roots" is also something you do once. Does Ancestry have subscriptions for that?

Wow, I never thought about how Voyager communicates with Earth. But now I wonder: if Voyager just sends photons towards the Earth, at the receiving end how are we recognizing which photons are coming from Voyager and how is the "signal" decoded?

I'd consider installing it if it had:

* In-depth technical explanation with architecture diagrams

* Open-source and self-hosted version

Also I didn't understand if it talks to a remote server or not. Because that's a big blocker for me.

That's only true now. You don't know that DNA leakage won't be a higher risk in the future (and FWIW, my opinion is the opposite of yours regarding the future risks). Moreover you can change your passwords, but you can't change your DNA.

> Even my brother and I share only ~50% DNA.

This is completely false. Any two random humans have more than 99% overlap by virtue of being the same species. It's even higher for brothers. We also share around 90% DNA with cats, dogs and elephants.

https://www.amacad.org/publication/unequal-nature-geneticist...

> I'm not too worried about it because it's never a 100% overlap.

This doesn't make sense. If they were equal, you'd be the same person except for environmental differences. Many applications don't need equal DNAs. E.g.

https://youtube.com/watch?v=KT18KJouHWg

> About insurance companies, they're legally forbidden to use such data.

This is a very weak argument. There's a long history of companies doing illegal things, and even if it's illegal today it doesn't mean it'll be illegal tomorrow.

They'd just bring it up shortly after under a different disguise.

Did you consider that your experience is but one of many? No need to be rude and bitter to the author.

The options might exist, but that doesn't mean that they are viable/usable. I agree with many of your points, but regarding document processing, LibreOffice really does not compare with any of the best alternatives (Google Documents, Office Word). It lacks options, if the options exists they are hidden in obscure menus, it doesn't display many things correctly (at least on my Linux computer), etc. For document processing, I'm still stuck on Google Docs, which is quite a sad state of affairs.

With Firefox you can toggle some settings that will make much harder to generate useful fingerprints. That's already a massive privacy difference.

Even if you split the search and ads divisions into two companies, what prevents them from entering into a partnership and keep doing whatever they do now?

Here we go again. Just require POST requests with a CSRF token for upvotes...

That's great if you don't want to understand what's going on under the hood, becoming the 2023 version of factory workers in Ford's assembly line.

Maybe the author shared the upvote link instead of the link to the post itself? It sounds like the most likely explanation.

I do wonder why this issue has never been fixed?!

A message ID only needs to be unique per workspace, in which case you'd expect very few collisions to begin with, and you could even retry on insert failure with a new timestamp. I don't think that would cause significant performance penalties.

These low-effort GPT comments should be removed as spam.