• AI Fire
  • Posts
  • 📱 #2 Meta AI App is “Laughably Bad”

📱 #2 Meta AI App is “Laughably Bad”

NotebookLM App Waitlist is Longgg

ai-fire-banner

Read time: 5 minutes

Meta’s new AI app just hit #2 on the iPhone Free app chart, but it’s not what you think. It feels more like a chaotic social feed, where sensitive infor can be saved, shared, and even tied to your profile… without you really knowing!

IN PARTNERSHIP WITH GENEVA TOURISM

Did you know that the World Wide Web was born in Geneva, Switzerland? Indeed, the first version of the Internet cropped up at CERN in 1989. Today the world-renowned center is home to the largest particle accelerator and to the CERN Science Gateway – a must-see hub for science enthusiasts that features hands-on exhibits, immersive virtual reality experiences, and live demonstrations.

AI INSIGHTS

nearly-6-10-workers-admit-making-non-transparent-ai-errors

In one of the most comprehensive studies to date - featuring 48,340 respondents across 47 countries and all global regions - a pattern has emerged in how employees use AI at work. Nearly half of employees admit they’re misusing AI, even when they know it breaks company policy.

🔓 The Reality of AI Use at Work:

  • 48 - 49% of AI-using employees have uploaded sensitive company data, including financial, sales, or customer information, into public AI tools.

  • Surprisingly, misuse is lower in companies without a policy (33%) or where employees don’t know if a policy exists (38%).

  • 47% have used AI in ways that could be seen as ethically inappropriate.

  • 57% reported using AI in non-transparent ways, such as:

    • Passing off AI-generated work as their own.

    • Not disclosing that AI assisted in completing a task.

📉 Work Quality and Over-Reliance:

  • 56% admit to making mistakes due to AI - often due to trusting inaccurate or “hallucinated” outputs.

  • 66% rely on AI outputs without verification.

  • 72% say they put less effort into tasks when using AI, indicating a rise in complacency.

In addition, more than 4 in 10 (44%) are “knowingly using it improperly” at work, but they still choose to ignore. This is a red flag. If we’re letting AI make decisions for us without really thinking it through, we’re setting ourselves, and our companies, up for trouble.

Why It Matters: We keep pushing adoption, but adoption without understanding is dangerous. We can’t just write a rulebook and expect everyone to follow it. We need ongoing training, examples, and a space to ask: Is this okay? Is this smart? We’re at a turning point.

TODAY IN AI

AI HIGHLIGHTS

📱 Meta AI app is now #2 on App Store. But it's actually a nightmarish social feed that saved & tied sensitive topics to the user profile. You can control its memory but it's “laughably bad”.

🚀 Anthropic just launched an AI for Science program using advanced models to boost high-impact scientific projects, by offering free API credits to researchers. Fill out the form to apply for Free!

💭 Cursor has come out with a list of the top AI models being used on its platform. Claude Sonnet 3.7 ranks #1.

top-ai-models-used-on-cursor

Which AI model do you use the most on Cursor?

Most useful and fastest one for you!

Login or Subscribe to participate in polls.

🆕 Google accidentally leaks details about its new Android design language, Material 3 Expressive. This is evolutionary, not revolutionary - an iteration of Material You in 2021.

🗣️ Google will launch a standalone NotebookLM app centered around its beloved, podcast-like AI Overviews feature. You can join the waitlist via the App Store or the Play Store.

💻 Nvidia just dropped Parakeet-TDT-v2, a super-fast, open-source speech AI that transcribes 1 hour of audio in 1 second. You can try it for Free here!

💰 AI Daily Fundraising: Cursor’s maker raised $900 million in funding, bringing its valuation to $9 billion. The company has raised over $173 million to date, with investors like Thrive Capital and a16z.

AI SOURCES FROM AI FIRE

It’s Monday again, and you know what that means! #5 Monday ToolDrop.

This week, I can say that I discovered some… “illegal” & mind-blowing AI tools like swapping your face in real-time with just a single click

Here’s what’s in store for you this week (Download a full version for FREE below):

  • Deep-Live-Cam

  • Cloudflare MCP Server

  • Free Programming Books

  • Suna – An open-source generalist AI agent

  • Developer Roadmap

Note: These exclusive resources & reviews are available only in our AI Fire community. It’s because you guys can freely ask for support or share personal experience during testing there. Get your full breakdown here (no hidden fee)!

ai-fire-academy

NEW EMPOWERED AI TOOLS

  1. 🎨 Midjourney’s Omni Reference easily inserts any character into a scene.

  2. 🎥 CrePal automates video creation for different platforms with AI agents.

  3. 🌊 Reef can analyze trends, create visuals, explain insights through podcasts.

  4. 🤖 Superexpert creates no-code AI agents to complete multiple tasks at once.

  5. Dash connects to your apps, remembers all contexts, and gets things done.

AI QUICK HITS

  1. 🚀 Microsoft’s cloud surged 33%, just outpacing Amazon’s AWS growth.

  2. 🩺 AI helped save a woman’s life, now cancer-free, following 30 years of smoking.

  3. 🔍 Pinterest updates its AI visual search, similar to Google’s multimodal one.

  4. 💡 Sam Altman hands Musk a win, OpenAI remains under non-profit control.

  5. ✍️ Over 250 CEOs sign open letter supporting K-12 AI: Microsoft, Etsy, Uber,…

AI CHEAT SHEET

AI CHART

voice-or-vice-chatgpt-anthropomorphic-tone-may-rewire-our-emotions

A study of 3 million ChatGPT chats & 4,000+ users highlights potential addiction patterns with its Advanced Voice Mode. Nearly 1,000 participants over 28 days, examining changes in emotional well-being across different usage modes (voice vs. text).

1. Emotional Engagement Isn’t Evenly Spread

  • Heavy users (the “long tail”) showed significantly more emotionally charged conversations. These users also reported a decline in emotional well-being by the end of the study.

2. Voice Mode Amplifies Affective Use, but Not Always

  • In natural use, voice interactions had more emotional cues than text. But in the controlled setting, voice alone didn’t increase emotional expression, users seemed to choose voice when seeking emotional connection.

  • Those starting with lower well-being improved more when using engaging voice; neutral voice with high usage made well-being worse.

3. Usage Time = Strongest Predictor

  • The amount of time spent with ChatGPT had more emotional impact than the mode (voice/text) or chatbot’s personality.

  • More time = more emotional cues, and sometimes more emotional strain.

The study introduces a critical new idea: socioaffective alignment: ensuring AI doesn’t just behave well logically, but also emotionally and socially.

It’s becoming harder to tell where support ends and dependence begins. Just like social media reshaped attention and self-esteem, voice-enabled AI could reshape how we manage emotions. But it’s a double-edged sword though.

AI JOBS

We read your emails, comments, and poll replies daily

How would you rate today’s newsletter?

Your feedback helps us create the best newsletter possible

Login or Subscribe to participate in polls.

Hit reply and say Hello – we'd love to hear from you!

Like what you're reading? Forward it to friends, and they can sign up here.

Cheers,
The AI Fire Team

Reply

or to participate.