A
@AutismCapital
That feeling when you recreate human stupidity for the sake of realism.
113
6
3.7K
@Polymarketposted on X
NEW: AI startup unveils "anti-Grammarly" tool that adds typos & grammatical errors to emails to make them seem human-written.
View original tweet on X →
Community Sentiment Analysis
Real-time analysis of public opinion and engagement
Sentiment Distribution
67% Engaged
49% Positive
18% Negative
Positive
49%
Negative
18%
Neutral
32%
Key Takeaways
What the community is saying — both sides
Supporting
1
Full‑circle irony:
We spent years building tools to make writing flawless, then built new tools to deliberately reintroduce mistakes so AI output looks human — people are amused and incredulous.
2
Market critique — a laundering economy:
Commenters see a circular SaaS stack (generation → detection → de‑detection) that monetizes plausibility, burns compute, and extracts margins without improving writing.
3
Signal inversion of authenticity:
Typos and small flaws have become trust signals; polished prose can now read like a bot, so imperfection is being engineered as a feature.
4
Escalating arms race:
AI writers, detectors, and anti‑detectors are already in competition — observers worry this will make it harder to tell humans from machines and create brittle detection dynamics.
5
Superficial fix:
Many argue adding typos is cosmetic — a “fake patina” that misses the real markers of human writing (quirks, priorities, tangents) and thus won’t truly replicate authenticity.
6
Practically unnecessary:
Several replies point out you can just prompt models to inject errors or manually remove em‑dashes; critics call the standalone product redundant.
7
Mixed adoption & consequences:
Some users find the idea useful for emails or cover letters, while others raise ethical and educational concerns about deception, trust erosion, and corporate communication implications.
Opposing
1
a prompt or simple feature anyone can run
(one person, turn off Grammarly, ChatGPT, iPhone AutoCorrect).
2
engineered authenticity erases the signal
and makes "real" indistinguishable from manufactured.
3
peak idiocracy
and a pointless use of attention and resources.
4
VC valuations and bubble behavior
or a shorting opportunity.
5
disrespectful to non‑native speakers
, with replies calling out xenophobic undertones.
6
gamed and degraded
, producing worse incentives long‑term.
7
AI to detect or undo
these manufactured signals so real authors can be verified.
8
intentional errors can themselves flag LLM use
, so the tactic may backfire.
9
real (bot/simulation)
rather than a serious startup.
10
correcting grammar isn’t inherently bad
, and some argue current AI can’t reliably mimic human messiness.
Top Reactions
Most popular replies, ranked by engagement
B
@BillyM2k
🤣 gotta remove em-dashes and weird ai-isms too
40
13
2.5K
D
@drive_dare
Next genius move: self driving cars that occasionally veer into ditches so cops think it’s a real Uber driver.
28
0
917
L
@LeifInvests
Dude we really went full circle 😂
25
0
965
T
@TrustWallet
what stage of the simulation are we in right now?
13
4
1.3K
Z
@zkespresso
Wtf we don't need a startup doing that, current AI is already allowing that, just prompt it
10
1
843
Report an Issue
Found something wrong with this article? Let us know and we'll look into it.