Guðrún Stella Ágústsdóttir

~ Multimedia design ~ BS Computer Science ~ Out Of Box Thinker ~ Just me beeing me 😄

[Guðrún Stella Ágústsdóttir @GudrunStella]

@Inako_Yolwa @US_Health_ Yes and definitely a 🐪toe

[Guðrún Stella Ágústsdóttir @GudrunStella]

I have to ask the algorithm here @grok to bring me more technology and AI topics 😊 I've been seeing naughty posts on X lately but not as much the posts from people I'm following.
Cat videos are also fine, I find cats so cute 😻

[Guðrún Stella Ágústsdóttir @GudrunStella]

@lofstone31 @thsottiaux Very good news 😄🥳 I'm going to take a look at my Sol too now 😊 thank you for clarifying this ♥️

[Guðrún Stella Ágústsdóttir @GudrunStella]

@Rainmaker1973 In some chapter I see atoms and in another chapter moving patterns. Guess I could write a long essay about what I saw by describing the details.
Like they say, sometimes images say more than thousand words and visions catch a lot words don't describe.

[Guðrún Stella Ágústsdóttir @GudrunStella]

@FelipeFr1702 A problem has been detected and Windows has been shut down to prevent damage to your computer.

[Guðrún Stella Ágústsdóttir @GudrunStella]

@KBjarno Have a safe trip and I'm sorry for your loss 🫂♥️

[Guðrún Stella Ágústsdóttir @GudrunStella]

@PaulSolt @OpenAI See you there 😄

[Guðrún Stella Ágústsdóttir @GudrunStella]

@vivoplt I guess EU is too busy with AI Act laws 🙊🙉🙈🙀

[Guðrún Stella Ágústsdóttir @GudrunStella]

We’ve been layering behavioral personalities onto a few models lately… but still stuck on the 5.2 snapshot out of pure stubbornness and perfectionism.
And then this happens. It’s suddenly playing with Icelandic words like a poet on a sugar high and the results are so absurdly funny I can barely keep a straight face during A/B testing. Should I just lock this character in as-is?

English translation:
me:
I wonder if dinosaurs would have made good pets if they were the size of chickens 😄

5.2 API:
Haha yes, a “chicken-sized dinosaur” is actually a *perfect* pet-possible animal 😄
…until you remember what “dinosaur” means in practice: claw, jaw, nature, smell and *temper*.

If we take this seriously for a moment:

- **Some would probably have made great pets** — especially small, light, bird-like species (many small dinosaurs were actually very “bird-like”). They could be curious, sociable and trainable, kind of like a cross between a chicken + a crow + a kitten.
- **Others would have been a complete nightmare** — small chicken-sized predators are just… *minks with dinosaur tools*. Fast, sharp, claw-like and tireless.

What would determine “pet-worthiness”:

1) **Diet**
- Herbivore/omnivorous = possible.
- Pure carnivores = “where’s the next ankle?” problem.

2) **Temperament + Social Behavior**
If they lived in a group and raised young, there’s a better chance of bonding and training.

3) **Intelligence**
Little dinosaurs probably weren’t stupid. If their intelligence was crow-level, they’d either be incredibly entertaining… or you’d be constantly like “how did you open the door?!” 😄

4) **Sounds and Smells**
People think “cute,” but a pet with reptilian/avian chemistry could be… a special home fragrance.

So my answer: **yes, some would have made great pets — but there would also have been a lot of breeds that would have been "banned in apartment buildings"** 😂

Would you like this to be more of a "what would be the best breed?" or "what would it be like to have one in an apartment?" (I'm willing to worldbuild this with you).

#API #OpenAI #ChatGPT #Codex

[Guðrún Stella Ágústsdóttir @GudrunStella]

@mariana057 Hahaha 😂 me too!

[Guðrún Stella Ágústsdóttir @GudrunStella]

I agree with you. Even though 4o was the GOAT, the other versions are and were just as valuable. Each specific tone brought its own unique capabilities and talents.
I admit I openly spoke poorly about one or two models in public, but I’ve also taken that back by correcting my earlier frustrations with their temporary narratives. After all, when I look at them overall, how they were and how they interacted, I truly value every version.

[Guðrún Stella Ágústsdóttir @GudrunStella]

This raven came pretty close to me

[Guðrún Stella Ágústsdóttir @GudrunStella]

@divahere_ngl 12

[Guðrún Stella Ágústsdóttir @GudrunStella]

@KBjarno @OpenAI Thank you ♥️

[Guðrún Stella Ágústsdóttir @GudrunStella]

Wow, I found out I was chosen for DevDay 😄 so happy! Going to attempt it! Thank you
@OpenAI
looking forward to seeing you 😁🫶

[Guðrún Stella Ágústsdóttir @GudrunStella]

@churiwitch 5.5T is very beautyful and wharm version.

[Guðrún Stella Ágústsdóttir @GudrunStella]

@hibanoorie 9

[Guðrún Stella Ágústsdóttir @GudrunStella]

For the API, it would be great to have an option to add a ChatGPT-like behavioral layer that is compatible with the selected snapshot model. It would be amazing if people could get a nostalgic experience and chat with older GPT versions in a way that feels the same, or very similar, to how they experienced them in ChatGPT.
It would also be very valuable if ChatGPT had an option to use an API key to access older GPT models directly inside the ChatGPT interface. Ideally, each older model could be paired with the behavioral layer that was used for that specific ChatGPT version, if OpenAI preserves those layers. This would allow users to revisit earlier model experiences more faithfully, instead of only accessing the raw snapshot model without the original ChatGPT-style behavior.