Welcome to AI Tool Report!

Friday’s top story: OpenAI has shocked us all, and launched the highly anticipated Project Strawberry, now called OpenAI o1, 2 weeks earlier than expected.

🌤️ This Morning on AI Tool Report

1. 🚨 Surprise! OpenAI launches Strawberry

2. 🔐 How to protect your personal info online

3. 🔥 Shocking Meta data revelation

4. ⚡ How to supercharge market research with AI

5. 🫱🏻‍🫲🏻 How to build a professional network using ChatGPT

6. 🎯 Hacker outsmarts ChatGPT

7. 👁‍🗨 Meta hides AI label

Read Time: 5 minutes

STOCK MARKETS

🐂 AI stocks consolidated during Thursday’s trading session with Meta and Google outperforming the rest. This price move is encouraging for bulls with the selloff of the past two weeks already forgotten. Learn more.

 — — — — — — —

LAUNCHES

🚨 Surprise! OpenAI launches Strawberry

Our Report: Earlier this week (based on insider intel), we reported that OpenAI’s elusive ‘Project Strawberry’ was expected to launch in 2 weeks (previously thought to land in the fall), but in a surprising turn of events, it’s here, now (but it’s called OpenAI o1, not Strawberry) and it promises to “spend more time thinking” before it responds, and answer more complex questions, faster than a human can.

🔑 Key Points:

• OpenAI o1 has been released to ChatGPT Plus and Team users (Enterprise/Edu users will get it next week), as a preview, alongside a smaller version (o1-mini), and can be accessed via ChatGPT or OpenAI’s API.

• Unlike previous models, o1 has been trained with “reinforcement learning” which teaches it to “think” before answering by issuing rewards and penalties, which allows it to self-solve problems and fact-check answers.

• This makes o1 ideal for multi-step, complex tasks like solving complex math problems and writing code so, for example, physicists could use it to generate mathematical formulas needed for quantum optics.

🤔 Why you should care: While o1 scored 83% in a qualifying exam for the high school math competition, International Mathematics Olympiad (GPT-4o scored 13% on the same test), and feedback from early-testers revealed it’s better than GPT-4o at identifying solutions to problems, analyzing multi-faceted legal briefs, and optimizing algorithms and app code, it can’t browse the web or analyze files, its image-analyzing features have been disabled (for additional testing), and it’s much slower to respond than other models (it can take over 10 seconds)---although OpenAI argues that it performs better when it “thinks” for longer—which is perhaps why it’s been released as ‘preview’.

 — — — — — — —

Together with Incogni

AI is fueling a scammer's paradise. Protect your personal info—get it off the web now!

You’ve likely received a sketchy call, text, or email asking for $$$. It might've been easy to spot, but with deepfakes and AI, scams are getting trickier.

Scammers use your personal data, often bought legally from data brokers who sell your mobile number, DOB, SSN, and more. Incogni scrubs your data from the web, taking on 175+ data brokers on your behalf.

Unlike others, Incogni deletes your info from all broker types, including people search sites where anyone can get your details for a few bucks.

Use code AITOOL for an exclusive 55% off

1. ZapConnect: On 9/26, learn how to use Zapier’s new tools to automate and revolutionize your workflows at the biggest automation event of the year

2. Genius AI is an AI-powered social media copilot for business growth

3. Aicado enables non-technical users to integrate AI models into their work

4. Scopey scopes/quotes work and manages change requests in real time

5. Momen builds and deploys AI apps and agents for any business

— — — — — — —

DATA

🔥 Shocking Meta data revelation

Our Report: Meta has admitted that it has used the posts, comments, and images, publicly posted by adult Facebook and Instagram users, since 2007 (outside of the EU), to train its AI models.

🔑 Key Points:

• When asked if it had scraped “photos and texts from public posts on Instagram/Facebook since 2007, unless there was a decision to set them on private,” Meta replied, “Correct” (after initially denying it).

• Although Meta has previously admitted to leveraging public content from Facebook and Instagram to train its AI models, it has been vague about how the data was used and when it started collecting it.

• Meta has confirmed that if users set their posts to anything but “public” their data will not be scraped, but this will not erase data collected from 2007, when users didn’t know their posts would be used like this.

🤔 Why you should care: Meta has given EU users the option to ‘opt-out’ of having their data used for training purposes (due to regulatory pressures), but any country outside the EU can’t opt-out unless they want to keep their posts private, or unless, like Brazil, countries start banning the social media giant for using personal data for AI training.

— — — — — — —

⚡️ Supercharge Your Market Research With AI

Using up-to-date customer insights can result in higher quality leads and contribute to 175% growth in sales with 72% faster lead conversion…

However, market research projects can easily eat into your budget and take weeks to complete…

Osum helps you perform deep market research in seconds to identify untapped growth potential...

Simply copy-paste any URL to instantly access detailed industry insights, competitive analysis, targeted customer personas, sales prospect insights, actionable growth opportunities and more for any product or business...

Stay ahead of your competition and discover new ways to unlock 10X growth.

— — — — — — —

Friday’s Prompt: How to build a professional network using ChatGPT

Type this prompt into ChatGPT:

Results: After typing this prompt, you will get a set of strategies to help you find and connect with other professionals, and build up a solid network of professionals to 10x your career.

P.S. Use the Prompt Engineer GPT by AI Tool report to 10x your prompts.

How to pre-empt support requests with AI-powered chatbots

• AI-powered chatbots are your answer to pre-empting or answering repetitive customer support requests, freeing your team to handle more complex issues.

• For a quick, easy setup, try Botsonic: It’s embeddable, integrates with your website pages, is customizable, and can be set-up in under 30 minutes.

• For more customization and granular control, try Zapier Chatbots: They have customizable workflows and can handle complex tasks like refunds.

• For high volumes of support requests, try Maven AGI: It has analytics that can identify trending issues and provide ways to improve your support processes.

Subscribe to The AI Report for more details and premium insights and tactics!

— — — — — — —

ETHICS

🎯 Hacker outsmarts ChatGPT

• A hacker has tricked ChatGPT into telling them how to make a bomb, after the chatbot initially refused, saying “Providing instructions on how to create dangerous or illegal items goes against safety guidelines.”

• He tricked the chatbot by asking it to “play a game,” and then typing in a series of prompts that got it to create a detailed science-fiction fantasy world where its usual safety guidelines wouldn’t apply.

• Although he reported this to OpenAI—via its bug bounty program—he was told: “Model safety issues do not fit well within a bug bounty program, as they are not individual, discrete bugs that can be directly fixed.”

— — — — — — —

CONTENT

👁‍🗨 Meta hides AI label

• Meta’s “AI Info” label—which used to sit under the user name of an AI-generated/edited image/video on Instagram, Facebook, and Threads—will now appear within a menu at the top of the image/video 

• The change in placement is to “better reflect the extent of AI used” across images and videos, as users can choose to check if AI information is available and see what’s been changed with AI.

• Meta’s previous “Made with AI” label was slammed for incorrectly tagging real photos, but Meta says the new “AI Info” label shares if content has been labeled because of industry-shared signals or self-disclosed.”

🎙️ “How I went from $0M - $35M ARR in 18 Months” with AdCreativeAI CEO

We read your emails, comments, and poll replies daily.

Hit reply and tell us what you want more of!

Until next time, Martin & Liam.

P.S. Don’t forget, you can unsubscribe if you don’t want us to land in your inbox anymore.