we don't even have a physical office anymore...

Nice, congrats!

Very nice, it’s way snappier and faster than most SPA I had to use recently

@U0N9SJHCH thanks, I spent a lot of time architecting it. Every initial page load is fully server-side rendered, so it is effectively a MPA that acts as a SPA when JS is enabled.

Ah cool, so you have rum on the serverside and then you hydrate?

Yup

Very cool! I will be using this for word lookups in the future I think. I love that kind of knowledge about words, and it feels snappy. Congratulations on the release! 🎉

Thanks @U0522J5HN. For word look-ups I personally recommend https://ordnet.dk/ddo which shares a great deal of data with DanNet (DSL, who is the entity behind http://ordnet.dk, is our partner). I think DDO is the best Danish dictionary.

DanNet has a more limited set of entries since the focus is on interword relations.

I usually use http://sproget.dk a lot, which does lookups in Den Danske Ordbog as well.

Congratulations on releasing 🙌

And it doesn't seem to mind the odd Canadian dropping by from time to time! ❤️

yes, after the UK decided to leave the EU the UK people here decided they wanted to stay.

A similar thing happened in #C050VDRHW - most of the activity went here

And we didn't even leave the EU!!!

Are you a stevedore?

@U04V5V0V4?? Don’t understand your question.

that looks like the docks, where stevedores are employed. It was meant humorously but unlike you, it failed to land :man-shrugging::skin-tone-3:

i see, I’m very tired, apologies. Been a long few days.

ha, enjoy the 🍵 and 🏡

I don't recognize them

it is @U0525KG62 .... this is a meme from 10 years ago though.

I'm older now

and apparently snowier

@U0AQ3HP9U have a look at some of the Euroclojure video's from Krakow (can't remember the year). There are quite a few references to @U0525KG62 in it. (and he wasn't even there!)

I'm glad that inside joke has died. All I needed to do was be a hermit for 8 years or so

We are glad to have you back... and apologies for resurrecting the meme. I won't do it again.

TIL

Dang. That's some inside stuff

Haven't had time to look sideways at a game since I got back from paternity leave 🙃 Maybe soon

I will soon be going on paternity leave again and I hope to get some time for it 😛

Mine arrived Friday night! Thank goodness I had a few hours that night but nothing more so far!

I‘m probably 6-7 hours in. It's a 10/10 for me so far, plays even better than BotW did. The world and the quests feel much more interconnected this time around.

That’s what I’ve heard online. I’ve just got to Hyrule (maybe 4hrs in) and so far it’s it’s more of the same for me. Which is great because I loved BotW, but I’m not floored yet.

The game uses the same overall template as BotW and a tutorial is always gonna feel like a tutorial, especially one that is structured the same as the BotW one. For me the game really started when I reached Hyrule.

I also thought the tutorial was a bit slow, but I think it's cause it's a retread of the Great Plateau.

Nice. I’m definitely enjoying it 😁

It’s the most I’ve consecutively played video games in about two years!

Same 😅

kinda creepy.

I think it’s an effect of trying to make Midjourney less creepy. The previous version did weird things. Even when not putting an extra leg on people it was always something there that was stuff for nightmares. Now there is much less of that, but it seems to be at the price of playing it super safe and narrowing down the space of what can be created. And then maybe there’s just one archetype for it when there’s a “21 year old woman” in the prompt. This is all pure speculation on my side. I don’t know anything about the stable diffusion architecture.

the fact that they are all white is also :man-shrugging::skin-tone-3:

It is a pretty/beautiful girl/woman, but as Ray said, they are all white. But you can see it is not a real photo. real skin looks very different I think and not as "perfect" (what ever that means)

Lots of photos in magazines are edited for things like that.

I wouldn’t read in too much in that all are white since it is all the same woman. While I was shrinking the prompt there was actually two grids that had black women.

> a close up photograph of a 21 year old woman dramatic and stunning award winning photo, dramatic linear delicacy, shot on Sony aiii high resolution digital camera, global illumination, TIME cover photo, f/11 --uplight --ar 2:3 --q 2 --style raw --v 5.1

> a close up photograph of a 21 year old woman dramatic and stunning award winning photo, delicacy, shot on Sony aiii high resolution digital camera, global illumination, TIME cover photo, f/11 --uplight --ar 2:3 --q 2 --style raw --v 5.1

Dunno why those doesn’t show up on my Midjourney profile. It’s quite arbritrary what does and what doesn’t.

ha, yeah maybe we shouldn't worry about AI bias (it's a joke - cos that's real) but even still, the variety of models and skin tone is low. I'm basically agreeing but please let's 🔪 🔫

( yes, I know you're a 🔪 sharpener! )

My AI analysis might not be super sharp. Using version 4 also seems to be using a very narrow selection of archetypes. A bit more variation, but it still seems that all 21 year old women are the same Irish river dancer.

They do look quite similar

v5 creates more photorealism, but is also even more boring than v4. 😃

The data bias is real, but the averaging effect itself also serves to make humans more beautiful. You can observe the same if you stack random people's faces on top of each other: it ends up looking like a model after a while.

The question is why it would average everything it has that hard regardless of the prompt. Did you see the original prompt? Prompt engineering deluxe. Then it creates the same image even with the prompt being just “Close up photo of 21 year old woman”. In the latter case I can understand that we get something generic. (Even if it’s strange it is all roughly the same image.)

That I can't explain :man-shrugging:

I agree that it is very odd.

It’s cool to know that we are beautiful on average. 😃