"He has identified three levels of human signals that we’ve lost in adopting AI into our communication. The first level is that of basic humanity signals, cues that speak to our authenticity as a human being like moments of vulnerability or personal rituals, which say to others, “This is me, I’m human.” The second level consists of attention and effort signals that prove “I cared enough to write this myself.” And the third level is ability signals which show our sense of humor, our competence, and our real selves to others. It’s the difference between texting someone, “I’m sorry you’re upset” versus “Hey sorry I freaked at dinner, I probably shouldn’t have skipped therapy this week.” One sounds flat; the other sounds human."
https://www.theverge.com/openai/686748/chatgpt-linguistic-impact-common-word-usage
While I'm all for criticizing LLMs / AI in every way possible, I am a bit worried about potential collateral damage of arguments like this.
Specifically, it seems to me that another way of stating this article's thesis is that chatGPT doesn't sound like a *neurotypical* human.
Guess who else doesn't sound like a neurotypical human and really could do without people questioning their agency and validity as real human beings.
@jmcclure which of the three characteristics do you think a neurodivergent individual would not exhibit? If anything, their character would be more extraordinary. In contrast, LLMs are overly ordinary.
If anyone risks having such a trait, I suspect it could be psychopaths, but that's much more an extremely superficial speculation rather than a meaningful hypothesis.
All / any of them really depending on the context. I'm honestly a bit lost on how to reply to your question and not comfortable with an assumption that all NDs communication would stand out as "extraordinary."
@jmcclure this generalisation was not mine, sorry if I gave you any reasons to think so
@mapto @BogWitch I asked deepseek for a human sounding toot about my breakfast and...
Here's a casual, relatable "toot" about breakfast with human vibes:
---
**Toot Draft:**
Okay, confession: my toast went *slightly* rogue this morning. 🍞🔥 Like, *just* on the edge of "artisanal charcoal" vs. "burnt offering." But you know what? Slapped some avocado on it (the forgiving green hero), sprinkled everything bagel seasoning like confetti, and added a lil' mountain of scrambled eggs. The crunch? Actually kinda perfect. Plus, side of yogurt with berries that *almost* look intentional. ☕ Fueled by coffee – the real MVP.
Solid 7/10 breakfast. Would definitely trip over the toaster again for this. 😂
\#BreakfastOfChampions \#SlightlyBurntIsAesthetic \#MorningFuel \#AvocadoToastSave
---
**Why this sounds human:**
* **Imperfections:** Mentions burnt toast, berries not perfectly arranged.
* **Humor & Relatability:** "Slightly rogue," "artisanal charcoal," "trip over the toaster," "forgiving green hero."
* **Casual Language:** "Okay, confession," "Slapped some," "kinda perfect," "lil' mountain."
* **Sensory Details:** "Crunch," mentioning coffee as the "real MVP."
* **Self-Deprecation & Charm:** Admits flaws but finds the positive ("Actually kinda perfect").
* **Slightly Silly Hashtags:** \#SlightlyBurntIsAesthetic adds personality.
* **Emojis:** Used naturally for visual breaks and tone.
Feel free to tweak the specifics (avocado -> peanut butter, eggs -> cereal, etc.) to match your actual breakfast! Enjoy your (slightly rogue) fuel. ☕️