Tim Kellogg tkellogg

Why “Stochastic Parrots” May Be the Death Knell for AI Progress

Written By: (not) Gary Marcus

Yann LeCun once said that “predicting the future of AI is like predicting the weather—you can get the next few days right, but everything beyond that is just elaborate guesswork.” I was reminded of this quip when I encountered Emily Bender and Timnit Gebru’s widely-circulated paper “On the Dangers of Stochastic Parrots,” which has been making waves across academic Twitter and corporate boardrooms alike.

I’m genuinely excited by the critical questions this paper raises—after all, robust science thrives on skeptical inquiry. Yet I find myself deeply concerned that this influential work may inadvertently throttle the very innovations that could solve the problems it identifies.

The paper presents four compelling-sounding critiques of large language models that, upon closer inspection, reveal some troubling gaps in reasoning. Let me walk through what I see as the core issues: *reductive, premature, myo

Model	Size	Min	Max
phi3.5	3b		0.85
phi4	14b	0.85	0.9
mistral	7b	0.98	0.98
llama3.1	8b	0.8	0.87
llama3.2:1b	1b	0.03	0.99
qwen2.5	7b	0.9	0.92
qwen2.5:14b	14b	0.92	0.98
qwq-patched	32b	0.65	0.75

Preface

This book started with this post from Ethan Mollick.

If you're not already aware, OpenAI researchers are known for pumping up the hype with vague implication-rich tweets about the AI that they're working on. I figured, Claude's good at this, why not use Claude to pretend for a little while. What if they weren't just hyping? What if they actually uncovered the mysteries of the cosmos?

The Origin of Ana's Powers: A Tale from Arendelle

It happened on her eighteenth birthday, during the first winter after the Great Thaw. Ana was walking through the castle gardens, her boots crunching in the fresh snow, when she noticed something peculiar about the roses. Despite the freezing temperatures, a single rosebud remained, perfectly preserved and waiting to bloom.

As she reached out to touch it, wondering if Elsa had somehow protected it with her ice magic, the flower suddenly burst into vibrant life beneath her fingers. Petals unfurled in brilliant crimson, and warmth radiated from where her skin made contact with the stem. The heat spread up her arm like liquid sunshine, and in that moment, Ana discovered her gift - the power of eternal spring.

Unlike her sister's ice powers, Ana's magic manifested as pure life force. Where Elsa brought winter's beauty, Ana could coax flowers to bloom in the deepest snow and warm the earth beneath her feet. The castle gardeners soon discovered that any plant A

in a room of 100 people, 99% are left handed. How many left handed people have to leave the room in order to bring that percentage down to 98%?

To solve this problem, let's break it down step by step.

Initial Number of Left-Handed People:
    Total people in the room = 100
    Percentage of left-handed people = 99%
    Number of left-handed people = 99% of 100 = 0.99 * 100 = 99

Goal:

Question: 99 Left-Handed People

in a room of 100 people, 99% are left handed. How many left handed people have to leave the room in order to bring that percentage down to 98%?

<|assistant|> Let's break down the problem step by step:

Initially, there are 100 people in the room, and 99% of them are left-handed. This means:

[ \text{Number of left-handed people} = 0.99 \times 100 = 99

From QwQ

Prompt

tell me a fart joke

Output

Uh, okay, I'll try to come up with a fart joke. Hmm, jokes about flatulence can be a bit tricky because they might not be appropriate in all situations, but if it's okay, I'll give it a shot.

	# /// script
	# requires-python = ">=3.12"
	# dependencies = [
	# "anyio",
	# "pydantic-ai-slim[mcp,streaming,code]",
	# "python-dotenv",
	# "mcp",
	# "typer",
	# ]
	# ///

	The Paper Trail: A Journey Through Contract Lifecycle Management
	Part I: The Awakening
	Maya Cohen brushed a strand of dark hair from her face as she stared at the mountain of paperwork on her desk. Three years out of law school, and here she was, the newest addition to the legal department at Horizon Industries, a mid-sized manufacturing company with global aspirations. The office hummed with the quiet intensity of Monday morning, the scent of coffee hanging in the air like a promise of productivity.
	"Good morning," came a voice from her doorway. Thomas Reed, General Counsel and her new boss, leaned against the frame. "Ready for your baptism by fire?"
	Maya nodded, trying to project a confidence she didn't entirely feel. "Absolutely."
	Thomas gestured to the stack of papers. "That's our current contract backlog. The company is growing faster than our processes. We need someone to overhaul our entire approach to contract lifecycle management. I think you're the right person for the job."
	"Contract lifecycle mana

	import sys

	def prepend_line_numbers():
	for line_number, line in enumerate(sys.stdin, start=1):
	# Remove trailing newline character if present
	stripped_line = line.rstrip('\n')
	print(f"{line_number}: {stripped_line}")

	if __name__ == "__main__":
	prepend_line_numbers()