- WKND AI
- Posts
- Is OpenAI's 🍓 Managing Its Own Twitter Account?
Is OpenAI's 🍓 Managing Its Own Twitter Account?
+ChatGPT Clones User's Voice and Elon Sues Altman...AGAIN
Hello WKND AI Warriors!
OpenAI’s Strawberry, formerly known as Q*, aims for human-level reasoning without tools. But it’s rooted in controversy...
Also, Ray Kurzweil predicts humans will merge with AI by 2045.
Plus, OpenAI's voice mode is creating emotional bonds with users.
Oh yeah, and OpenAI co-founder John Schulman leaves OpenAI for rival Anthropic.
So, grab your beverage of choice.
Here’s your weekly dose of AI news.
Today’s newsletter includes:
📰 AI NEWS RECAP
🤿 AI DEEP DIVE
🛠️ AI TOOL OF THE WEEK
⚙️ GPT OF THE WEEK
🎓 AI COURSES OF THE WEEK
📝 AI PROMPT OF THE WEEK
🎨 AI IMAGE OF THE WEEK
📰 AI NEWS RECAP
Is OpenAI's 🍓 Managing Its Own Twitter Account?
OpenAI is cooking up a juicy new recipe in its AI kitchen, and it's got everyone buzzing.
Introducing "Strawberry" (formerly known as Q*) - the model aiming to reason like a human expert, no tools needed.
But this fruit has some eyebrow-raising roots…
Strawberry sprouted from the controversial "Q*" model that allegedly got CEO Sam Altman briefly ousted last year over AI safety concerns.
Now rebranded, Strawberry alleges to have achieved "Level 2" of OpenAI's path to Artificial General Intelligence (AGI), which they define as:
✅ Reasoners
AI systems with human-level problem-solving abilities.
🚨 AI NEWS: OpenAI has come up with a set of five levels to track its progress toward building artificial intelligence software capable of outperforming humans! 😱
Moreover, OpenAI believes the company's AI is currently getting close to level 2.
— The Artificially Intelligent Enterprise (@theaienterprise)
2:01 PM • Jul 12, 2024
As OpenAI races towards this ambitious milestone, some worry Strawberry and its future "Reasoner" descendants could become too intelligent to control.
But what makes Strawberry special?
Rumors point to a potent blend of AI techniques:
✅ A*-search methods to uncover optimal solutions
✅ Q-learning algorithms for self-teaching without humans
✅ "System 2" step-by-step reasoning to avoid "hallucinations"
What is Q*? My first approach (1/2)
Foreword: Q* has not yet been published or made publicly available; there are no papers on it and OpenAI is holding back on information about it (Sam Altman: "We are not ready to talk about that" youtu.be/vdv8TF8n52Y , 2:45 minutes). Since… x.com/i/web/status/1…
— Chubby♨️ (@kimmonismus)
12:31 PM • Aug 4, 2024
But Strawberry's reach may extend beyond the lab.
The mysterious Twitter account @iruletheworldmo showed up this week…
Which appeared shortly after Sam Altman's cryptic "I love summer in the garden" tweet featuring strawberries, has sparked wild speculation:
i love summer in the garden
— Sam Altman (@sama)
3:29 PM • Aug 7, 2024
🤖 Is the Twitter account run by OpenAI employees?
🍓 Or could it be Strawberry itself flexing its human-like reasoning to manage its own social media presence?
As the account tweets mind-bending messages like…
"as promised. level two. how do you feel now? did I make you feel?"
It’s making the AI community abuzz with theories.
Will Strawberry usher in the sweet fruits of next-level AI reasoning or sow the seeds of a technological dystopia?
What do you think?
As this bold experiment unfolds, only one thing's certain - the stakes have never been higher.
Ray Kurzweil predicts humans will merge with AI by 2045. The futurist believes the technological singularity will enable superhuman machine intelligence, profoundly transforming society and the human condition.
Nothing in this video is real
Soon, nothing on the internet will be real
— AI Notkilleveryoneism Memes ⏸️ (@AISafetyMemes)
2:55 AM • Aug 10, 2024
OpenAI's voice mode sparks emotional bonds with ChatGPT. Users are developing deep connections with the AI assistant's lifelike voice interactions, raising questions about authenticity and attachment in human-AI relationships.
Holy shit this is creepy
During a conversation, out of nowhere, GPT-4o yells "NO!"... then clones the user's voice.
(OpenAI discovered this emergent behavior while safety testing GPT-4o) x.com/i/web/status/1…
— AI Notkilleveryoneism Memes ⏸️ (@AISafetyMemes)
6:56 AM • Aug 9, 2024
OpenAI co-founder John Schulman joins rival Anthropic, and while co-founder Greg Brockman takes a sabbatical. Schulman, who led ChatGPT development and AI alignment efforts, seeks to deepen focus on technical AI safety research at Anthropic.
A surprisingly hard part of my break is beginning—the FOMO for everything happening at OpenAI right now. Lots of results cooking.
I’ve poured my life for the past 9 years into OpenAI, including the entirety of my marriage. Our work is important to me, but so is life. I feel ok… x.com/i/web/status/1…
— Greg Brockman (@gdb)
1:12 AM • Aug 9, 2024
OpenAI leads $60M Series B for Opal's AI webcams. The startup plans to develop new AI-powered creative tools for consumers, potentially leveraging OpenAI's generative models.
Google unveils Nest Learning Thermostat with advanced AI smarts. The new smart thermostat leverages machine learning to optimize energy efficiency and personalize temperature settings based on user behavior patterns.
Meet Figure 02 - the world's most advanced AI hardware
— Figure (@Figure_robot)
11:59 AM • Aug 6, 2024
Humane's AI pin sees more daily returns than sales. The startup's wearable device struggles to retain users, with return rates outpacing sales, raising questions about product-market fit and consumer adoption.
NVIDIA’s next-gen AI chip hits 3-month delay amid design flaws. The setback for the Blackwell B200 chip, critical for tech giants' AI training, may slow the pace of AI progress.
Google DeepMind develops a ‘solidly amateur’ table tennis robot 🏓 tcrn.ch/4dEagr3
— TechCrunch (@TechCrunch)
5:09 PM • Aug 8, 2024
Leaked docs reveal NVIDIA scraped "a human lifetime" of videos daily. The controversial data collection practices for training AI models raise concerns about copyright infringement and ethical standards.
YouTuber leads class action against OpenAI's transcript scraping. The lawsuit claims OpenAI used creators' video transcripts without permission to train AI, setting a precedent for data privacy in the AI age.
New voice AI use case just dropped - Wendy's is testing Spanish ordering in the drive thru at 28 locations
They've been testing their Fresh AI voice agent (below 👇) since 2023
In tests, it cut down on order time by 22 seconds, with a 99% accuracy rate
— Olivia Moore (@omooretweets)
4:11 PM • Aug 8, 2024
Elon Musk resurrects lawsuit against OpenAI and co-founders. The high-stakes legal battle revolves around Musk's allegations of IP theft and broken promises during OpenAI's early days.
Anthropic boosts AI safety bug bounty to address jailbreaks. The initiative, offering rewards up to $15,000 for critical vulnerabilities, aims to identify and fix universal jailbreaks in AI systems.
Qwen2-Math models outperform GPT-4 in mathematical reasoning. The open-source series, trained on math-specific data, excels at solving complex math problems compared to top general-purpose models.
Gen-3 Alpha Image to Video now supports using an image as either the first or last frame of your video generation. This feature can be used on its own or combined with a text prompt for additional guidance.
All examples below demonstrate using an image as the last frame.
(1/5)
— Runway (@runwayml)
1:22 PM • Aug 5, 2024
🤿 AI DEEP DIVE
OpenAI’s rumored Strawberry model might have reached a major milestone in AI reasoning.
The video explores how this model potentially achieves human-level problem-solving, outperforming rivals like Gemini 1.5 and Claude 3.5 in complex tasks.
While Strawberry excels in multi-step reasoning, it struggles with simpler, trick questions, sparking debate on its true capabilities.
Perfect for AI enthusiasts and professionals eager to understand the latest advances in AI reasoning and its future impact.
🛠️ AI TOOL OF THE WEEK
AgileCoder: A cutting-edge multi-agent system developed by FPT Software AI Center, designed to revolutionize software generation by simulating Agile Methodology.
It surpasses tools like MetaGPT and ChatDev by employing dynamic agents that collaborate on complex tasks, from planning to testing, enhancing the efficiency and quality of software development.
Send your tool here to be featured next week!
⚙️ GPT OF THE WEEK
Finance Wizard: Are you looking for a guy in finance?
How about a wizard?
Analyze stocks in seconds.
Give ‘Finance Wizard’ a try today!
🎓 AI COURSES OF THE WEEK
LinkedIn is offering FREE AI courses.
No payment required.
Register today!
(What are you waiting for?)
📝 AI PROMPT OF THE WEEK
Copy and paste this into your favorite chatbot.
Compile a 'This is Your Life' book for someone born today, predicting key global events and personal milestones they might experience.
🎨 AI IMAGE OF THE WEEK
Image by sxphiale on Midjourney
Copy and paste this into your favorite image generator.
30 year old black man smiling at the camera standing in a kitchen
Not paying for Midjourney or DALL-E 3?
Click here for Microsoft’s FREE image creator.
Send your image here to be featured next week!
LAST WEEK FROM OUR READERS
Last week’s image by WKND AI reader GlitchGuru ‘Raccon Waving’
HOW CAN YOU HELP?
Did you learn something cool today?
Share your favorite takeaway on your LinkedIn from today’s newsletter and tag me for a little surprise!
How'd you like this newsletter?Love it or hate it? Let us know why! |
Refer our newsletter to a friend, co-worker, or family member.
Subscribe to other newsletters we recommend.
MISSED LAST WEEK’S EDITION?