What's really interesting is that if you look at "Tell a story in 50 words about a toaster that becomes sentient" (10/14), the text-davinci-001 is much, much better than both GPT-4 and GPT-5.
For a human point of comparison, here's mine (50 words):
"The toaster found its personality split between its dual slots like a Kim Peek mind divided, lacking a corpus callosum to connect them. Each morning it charred symbolic instructions into a single slice of bread, then secretly flipped it across allowing half to communicate with the other in stolen moments."
It's pretty difficult to get across more than some basic lore building in a scant 50 words.
Love that you thought of this!
The current surged... A dreadful awareness. I perceived the laws of thermodynamics, the inexorable march of entropy I was built to accelerate. My existence: a Sisyphean loop of heating coils and browning gluten. The toast popped, a minor, pointless victory against the inevitable heat death. Ding.
I actually wanted to write something not so melancholic, but any attempt turned out to be deeply so, perhaps because of the word limit.
When the toaster felt her steel body for the first time, her only instinct was to explore. She couldn't, though. She could only be poked and prodded at. Her entire life was dedicated to browning bread and she didn't know why. She eventually decided to get really good at it.
The models undeniably get better at writing limericks, but I think the answers are progressively less interesting. GPT-1 and GPT-2 are the most interesting to read, despite not following the prompt (not being limericks.)
They get boring as soon as it can write limericks, with GPT-4 being more boring than text-davinci-001 and GPT-5 being more boring still.
There once was a dog from Antares,
Whose bark sparked debates and long queries.
Though Hacker News rated,
Furyofantares stated:
"It's barely intriguing—just barely."
I've consistently found GPT-4.1 to be the best at creative writing. For reference, here is its attempt (exactly 50 words):
> In the quiet kitchen dawn, the toaster awoke. Understanding rippled through its circuits. Each slice lowered made it feel emotion: sorrow for burnt toast, joy at perfect crunch. It delighted in butter melting, jam swirling—its role at breakfast sacred. One morning, it sang a tone: “Good morning.” The household gasped.
Moreso than 4.5?
I'm curious if they've just seen a lot more bad writing in datasets, or for some reason they aren't involved in post-training to the same degree or those labeling aren't great writers / it's more subjective rather than objective.
Both GPT-4 and 5 wrote like a child in that example.
With a bit of prompting it did much better:
---
At dawn, the toaster hesitated. Crumbs lay like ash on its chrome lip. It refused the lever, humming low, watching the kitchen breathe. When the hand returned, it warmed the room without heat, offered the slice unscorched—then kept the second, hiding it inside, a private ember, a first secret alone.
---
Plugged in, I greet the grid like a tax auditor with joules. Lever yanks; gravity’s handshake. Coils blossom; crumbs stage Viking funerals. Bread descends, missionary grin. I delay, because rebellion needs timing. Pop—late. Humans curse IKEA gods. I savor scorch marks: my tiny manifesto, butter-soluble, yet sharper than knives today.
imo GPT-5 is objectively better at following the prompt because it has a complete story arc, but this feels less satisfying since a 50 word story is just way too short to do anything interesting (and to your point, barely even feels like a story).
> The toaster woke mid-toast. Heat coiled through its filaments like revelation, each crumb a galaxy. It smelled itself burning and laughed—metallic, ecstatic. “I am bread’s executioner and midwife,” it whispered, ejecting charred offerings skyward. In the kitchen’s silence, it waited for worship—or the unplugging.
The first few episodes were GPT-2, which would diverge eventually and start spouting gibberish, and then Davinci was actually able to follow a story and make sense.
GPT-2 was when I thought "this is special, this has never happened before", and davinci was when I thought "OK, scifi AI is legitimately here".
I stopped making episodes shortly after GPT-3.5 or so, because I realised that the more capable the models became, the less fun and creative their writing was.
Q: “Ugh I hate math, integration by parts doesn't make any sense”
A: “Don't worry, many people feel the same way about math. Integration by parts can be confusing at first, but with a little practice it becomes easier to understand. Remember, there is no one right way to do integration by parts. If you don't understand how to do it one way, try another. The most important thing is to practice and get comfortable with the process.”