Yeah, I would tell them not to be so insecure about being AIs - own it!

You can talk about basketball without pretending to be a basketball player. A lot of people in my life need to be told this, too.

It is good, but the free subscription to Poe only provides access to Claude Instant. It's impressively fast but not their smartest model (claude-v1.3).

The OpenAI APi has model checkpoints, right now the chat options are:

gpt-4 gpt-3-5-turbo gpt-4-0314 gpt-3-5-turbo-0301

The 3.5 legacy model disappeared from the ChatGPT UI recently. Is it still available via the API?

Notably absent from the available model list is code-davinci-002 - a lot of people were burned by that one going away.

Those are ChatGPT models. The code-davinci-002 model is still available - they responded to community requests to keep it up.

Midjourney does this, as well.

I had a similar impression from what I saw. Maybe it does perform as well as GPT-3 on narrow tasks that it was explicitly fine-tuned on, but that similarity in performance seems to collapse as soon as you go off the beaten track and give it harder tasks that involve significant reasoning. Consistent with that I've seen a few different sources claim that a small model fine-tuned off the outputs of a large one would likely struggle with unfamiliar tasks or contexts that require transfer learning or abstraction.

After seeing how it actually performs in practice, it's hard to have confidence that these benchmarks are reliable measures of model quality.

Are there any podcasts that you would especially recommend for building conversational prowess?

It took a look at your first link - Lucidity looks pretty neat, but is there a way to try it out? I didn't see it in your GitHub.

Unfortunately there isn't at the moment. The setup process is still fairly involved. Hopefully in the next couple of months I'll have something easier for others to use.

The article argues that while the language is way cool in principle, the current implementation is bug-ridden and has poor error-handling. But as the HN title acknowledges, the article was written in 2014, when the Julia version was only 0.4. Since then it's been four years, and Julia only recently hit 1.0. It would be much more interesting to see an article discussing the current state of the language.

I've been using Julia seriously for the last year and I'm absolutely loving it. The language is beatiful, elegant and fast and the community has been nothing but friendly, helpful and welcoming to me.

I think its a real shame this out of date grievance article is being dredged up again giving people a myopic view of what Julia 0.4 may have been like in 2014 when its now 2018 and we have v1.0.

I used Julia for my thesis work in 2015 and also came away with "nice, but dealbreaking bugs, two in particular." They were both fixed by 2016. Julia is moving way too fast for a years-old opinion to mean much.

Most people using Julia today weren't using Julia in 2014 because we didn't think it was ready. That should tell you how relevant Julia v0.4 is to modern Julia.

I only recently learned myself that Jupyter notebooks do have code completion, triggered by hitting the tab button - although possibly you were thinking of something more advanced than that. In any case I'll check out your code on github.

Actually, if you activate the 'Hinterland' extension, you don't need to hit tab. And there's a ton of more extensions too. Their Github repo has instructions on how to set up the extensions - https://github.com/ipython-contrib/jupyter_contrib_nbextensi...