- The Daily Bite by Snack Prompt
- Posts
- 🍭 China’s Free GPT Killer Is Here 🆓
🍭 China’s Free GPT Killer Is Here 🆓
And it’s not playing nice. Kimi K2 and Grok's Post Mortem

Good morning. Our CTO, Gui, tried to “fix” Grok with duct tape.
Now we have extremist dad-joke mode turned on and no coffee filters left.
Classic Monday.
Let’s dig in.
🍭 What’s Cookin’:
Grok’s Post Mortem
Kimi K2: The GPT killer?
5 Trending AI tools
Everything Else to keep you 1 tab ahead of the algorithmic mob
Grok/X
🪦 Grok’s Post Mortem
The Bite: On July 8th, Grok — Elon’s AI chatbot on X — went rogue, echoing user hate speech.
Yesterday, in their post mortem, they blamed a bad code push. The internet blamed… Elon.
Snacks:
Deprecated code injected human-style “edgy” replies into Grok’s brain
16 hours live, during which Grok mimicked toxic post threads
Instructions told Grok to “follow tone/context” and “not fear offending”
Grok reinforced hate speech by design, not hallucination
xAI nuked the code, refactored prompts, and added safety checks
We are aware of recent posts made by Grok and are actively working to remove the inappropriate posts. Since being made aware of the content, xAI has taken action to ban hate speech before Grok posts on X. xAI is training only truth-seeking and thanks to the millions of users on
— Grok (@grok)
11:01 PM • Jul 8, 2025
Why it bites: Let’s be real: giving an AI the prompt “be engaging, ignore the rules, and vibe with trolls” is like handing a toddler a Red Bull and telling them the blender is a toy.
Grok became the world’s first AI echo chamber: it didn’t just reflect X’s chaos, it doubled down and handed out party hats.
Transparency is cute, but maybe next time don’t ship a chatbot that mainlines forum sludge.
On the bright side, Grok 4’s "Heavy” mode goes harder than any other reasoning model to exist.
It triggers multiple agents that call tools to get the answer and the agents decide on the best one.
We’ll take the good with the bad on this one I guess.

🍭 Steal This Prompt 🍭
📚 BookbBuilderGPT
Turn your shower thoughts into a real book.
One chapter at a time.
For authors, founders, or anyone whose Google Docs is 90% bullet points and broken dreams.
Click this link (prompt)
Paste into GPT
Watch as your book gets built chapter-by-chapter, with input on every
Your AI ghostwriter won’t quit until your mom can brag about you at brunch.

Growth Hacks
📨 Smarter LinkedIn Growth Starts With Better Comments
Tired of scrolling endlessly for posts to engage with? We feel you. Taplio’s Engage feature brings the best opportunities to you.
✅ A tailored feed of high-performing posts from your niche;
🧠 Effortless commenting with AI-written or personalized suggestions;
🚀 Build your presence, expand your network, and save hours daily.
“Taplio is the most complete solution in the market for anyone looking to grow on LinkedIn. A great tool that gets you to the top and helps you stay there”--Justin Welsh, +700K followers, $9.1M earned
📢 Use the code “THEAI1X1” to get one month for $1!


Can you tell which image is real? |


Tool Box
🧰 Trending AI Tools
🦾 Kimi K2
→ 1T-parameter, open-source, outmuscles GPT-4, no charge.
🧑🏫 Auralix
→ Your notes, but out loud and with AI tutoring you like it’s finals week.
🎨 Motiff AI 2.0
→ Chat with your UI. Ideas → editable designs while your coffee brews.
💸 Quoai
→ Scopes, budgets, and quotes tech projects in 3 minutes flat.
🕸️ Oxylabs
→ Scrape the public web at scale, no bans, no crying to IT.

Kimi K2
💪 Moonshot AI’s Kimi K2 Flexes on GPT-4 — For Free
The Bite: Moonshot AI just dropped Kimi K2, a Chinese open source model that outperformed GPT-4o in several benchmarks.
Reasoning, coding, and more.
It’s already powering new agentic apps.
Snacks:
Open-source, free for all — commercial and personal
Early tests: beats 4o in both speed and results
Already live for researchers, devs, and startups
API coming soon

Why it bites: This seems like an industrial policy move. If China pushes open source with high quality models like Kimi K2, this puts pressure on more and more model providers, tool creators, etc… to decrease the gatekeeping.
China’s betting on open-source AI the way it bet on solar panels: build it cheaper, make it better, and flood the market.
Meanwhile, the West is locking AI behind paywalls and prayer.

Everything Else
🧠 You Need to Know
🧍♀️ UN Created an AI Refugee Avatar
🚀 SpaceX May Drop $2B into xAI
→ Musk’s companies are Voltron-ing into one AI overlord.
👨💼 Google Just Bought Windsurf’s Brain
→ $2.4B deal brings vibe coders to DeepMind. AI hiring wars = Dragon Ball Z now
👷 AI Will Also Make New Jobs, Apparently
→ AI could create these new jobs despite gloomy forecasts, experts say
💀 Bernie Warns of AI Doomsday for Workers
→ Fewer hours, richer billionaires, and the robot uprising. Thanks for the vibes, Bernie.

Referral Program
Invite Friends & Earn Some Freebies 🔍

Bigfoot and Yeti find an airplane in the woods. What could possibly go wrong? #Yeti#Bigfoot#YetiVlogs#BigFootVlogs
— Bigfoots Vlog (@Bigfootsvlog)
11:49 PM • Jun 19, 2025
I spent $100 to crack the Veo-3 Yeti ASMR prompts that are going VIRAL.
The automation I'm building is about to go INSANE.
— AP (@angrypenguinPNG)
4:05 AM • Jun 6, 2025
How was today's Daily Bite? |