Hedepig

I think we're very used to encountering projects that have a level of polish or completeness. It's jarring when we come across amateur (for the love of) projects and homepages that lack such.

For the record, I agree, I found I wanted a better look at all the theme had to offer.

I am just a little fed up with this kind of entitlement that comes when someone puts something they've created out there. We want to encourage people to release their stuff, even if it is a little rough. Give them tips, help them. Don't just attack it. IMO HN will be better for it.

But maybe that's just me

Also in its defence: it costs nothing.

Feedback and critisism can be great, but why be hostile towards projects where a person has used their time to build something and release it gratis.

> LLMs always talk in the same confident tone,...

They do, but they are better at verifying what they've already said. So a simple prompt asking them to verify the facts they've presented often improves the accuracy. There are also other techniques like chain of thought and tree of thought that further improves accuracy.

> at least some of the time

YMMV

The recent additions OpenAI have made allows for tighter control over the outputs. I think that is a very useful step forward.

I have managed to get doom emacs working on termux. I presume the same is possible with vim.

Somewhat tangimential, are there any open source attempts to compete with OpenAI's embeddings?

I know Word2Vec is a thing but I believe that is on a word by word basis, and doesn't capture the semantic meaning of whole sentences and paragraphs.

They charge so little for embeddings I secretly hope they do open source it. Because if for some reason it is stopped, any search functionality or the like that relies upon the API would cease to function

While this technology is becoming mainstream, those of us that are okay with it should understand there is something fundamentally "icky" for some people. Which is difficult to debate rationally.

This is partly the fault of Sci-fi, bio-engineering is rarely shown as a positive in TV and Film.

This will be mostly solved with time, eventually it may become economically unviable to buy "field meat" (not sure what the term will be), and it will become a lot more accepted when they see the people around them enjoying it without severe consequences.

With the right equipment, if not now, soon you will be able to.

Maybe, but really we don't know what the shape of the curve until we actually find the limit.

I think you're right, but history has plenty of technological dead-ends, particularly in the AI/ML world

Emacs can be anything you want if you build it or what you want is built.

NAL but doesn't a defemation case require proof of malice?

We'll probably have tiered AI systems.

Upgrade to the Argue Bot 10 series and save big! With a 3x higher success rate in securing cheaper deals compared to the 1000 series, it's the smart choice for cost-savvy individuals. For an additional $999 per month, you'll have access to cutting-edge negotiation technology that will work tirelessly on your behalf. Don't miss out on potential savings - make the switch to the Argue Bot 10 series today!

Edit, changed it to an improved promo using ChatGPT from my original which is here: "Argue bot 10 series, 3x more likely to get you a cheaper deal than the 1000 series, for only an additional $999 per month"

> think around 4K words

4k _tokens_, 1 token is approx. 3/4 a word. So it's roughly 3k words

32k _tokens_. There are approx. 3/4 tokens per word, so that's around 24k words

This looks good. Is the training data in the repo itself?

The terms of ChatGPT/GPT-3.5 explicitly state that one cannot use their data to construct competitive models

Slightly tangential, is there some kind of crowdsourced effort to build training data for fine tuning? Alpaca used the training data built from gpt-3.5, so there are terms of use restrictions

It brings to my mind the levels of self driving

We're definitely at 2 right now, and picking away at level 3.

I have heard some people skeptical that we can overcome the problems of truthfulness due to the inherent limitations of LLMs. But, at least on the face of it, it appears we can make incremental improvements.

If only they would actually be OpenAI

I have seen

> for me

I question the objectivity of these percentages

Despite productivity generally improving over the last few decades, wage compensation has not.

I'm concerned this will continue as a trend with any productivity improvements from these models.

I don't think that AGI is binary. It's a (very difficult to measure) scale, the tech that OpenAI produced has brought us further along that scale in my opinion.

Despite all the (legitimate) criticisms, I still don't understand how some people are not impressed by what has been achieved by OpenAI.

I agree, it's far from perfect, but as a tool it's certainly usable and useful, with quite fascinating emergent behaviour.

I think they could have worded that better

Are you saying you're not impressed by what we have acomlished so far?

Do you assume it will stay the way it is?

Do you believe we're at the peak of what will be capable?

As shown by this post, we're still discovering what's possible with the tools we have now.

Also, human programmers aren't amazing at writing defect free code.

I think most people don't truly grasp how much this will change the world. IMO the advances in Synbio will be vastly more influential than the advances in AI (though AI is utilized heavily for the former)

In their defence they openly state its not for the final product but rather to get _something_ shipped.

Reducing cognative load is still entirely helpful

I predict AI assisted archicture design will come soon

> I still find coding a mentally exhausting task and maybe even more now that I am getting older

If it helps, we're at the dawn of AI assisted development. That will take a lot of the cognitive burden, allowing us to focus on the "what" more than the "how"

Not to be confused with Ook![1]

[1][https://esolangs.org/wiki/Ook!]