Discover the Power of Accurate Fact-Checking

Hey there! 🚀 The diligent team at has been burning the midnight oil to bring you our game-changing Fact-Checking Tool, which is now in its exciting BETA stage. And guess what? We’ve got some intriguing results from our recent study to spill the beans on!

Here’s the Scoop on Our Findings:

  • We pitted 6 AI heavyweights in the fact-checking arena:, GPT-4, GPT-3.5, CodeLlama-34b, Llama-2-13b, and Llama-2-70b.
  • Our testing battleground? A dataset bursting with 120 fresh-off-the-oven facts (60 truths and 60 fibs).
  • Leading the pack,’s Fact-Checking Tool hit a stunning accuracy score of 72.3%! GPT-4 wasn’t too far behind, clinching the silver spot at 64.9%.
  • was the star player with a 0% “unknown” rate and a minimal error footprint, showcasing top-tier reliability.
  • When put to the ultimate test of sifting through AI-generated articles, outperformed both GPT-4 and Llama-70b in spotting those sneaky false facts.

Why Every Click Counts: The Significance of Fact-Checking

In our AI-driven digital era, a tide of AI-produced content is swamping the internet. But here’s the kicker: the credibility of this ocean of information is under siege by a phenomenon known as LLM Hallucinations. What’s that, you ask? “Hallucinations” are those bold-as-brass AI responses that seem to pop out of thin air, without any grounding in their training data. And ever since chatGPT burst onto the scene, these hallucinations have been sneaking into online content at an alarming rate. The fallout? A potential wildfire of misinformation, with profound impacts on society.

The Shield: Our Fact-Checking Tool to the Rescue

Crafted with Content Editors in mind, our Fact-Checking Tool is your trusty sidekick in the quest for truth. Its mission? To drastically slash the chances of these hallucinations creeping into your content.

What’s under the hood? A sleek user interface boasting reference sources and enlightening explanations. These gems help users decode the reasoning behind flags on their statements, steering them towards dependable, top-shelf sources. 👇

See our Fact- Checking Tool at work: 

Join Our Beta Brigade! 🌟

So far, we’ve rolled out the red carpet for 40 lucky beta users, and the list is growing every day! Fancy a slice of the beta action? If you haven’t thrown your hat in the ring yet, there’s still time to hop on board!

📌 Just a heads-up: Make sure to apply using the email tied to your account. 👇 

clear mannequin on dark blue background

GPT4 Picture Interpretation Skills Are Incredible!

Me: Is the attached picture funny?ChatGPT

ChatGPT: Humor is subjective, so what one person finds funny, another might not. The attached picture seems to be a playful take on a photo editing request, where instead of just removing the person in the background as requested, the entire background has been changed. Some people might find this humorous due to the unexpected twist and the playful nature of the edit. Would you like more information or analysis on the image?

Me: yes, do you understand why the plant were dried in right picture? This is the reason the picture is funnyChatGPT

ChatGPT: Yes, I understand. The humor in the image arises from the literal interpretation of the request. When asked to “remove the guy watering plants in the background,” instead of just removing the man, the editor chose to depict what might happen if the man was never there to water the plants in the first place. Thus, the plants appear dried and withered in the second picture. It’s a play on words and expectations, and the unexpected result creates the comedic effect.

I asked ChatGPT if he understands (note that wrote “he” not “it”, I’m getting worried!) why the above picture is funny and he firstly replied according to the fact that my question could have been hiding a second sense, which is already quite surprising. But, when I asked about why the picture is funny, he exactly replied that it is because while removing the man, the editor also made the plants appear dried… that’s crazy, issn’t it? I mean if this is not yet AGI, but this is at least a glimpse of it common sense!

That interaction with ChatGPT really blew my mind and made me realize just how far AI has come. The ability to understand subtext and explain complex concepts is an enormous leap towards artificial general intelligence.

As an SEO expert and blogger, I’m fascinated by the implications of this technology. When I asked ChatGPT if it understood why the meme was funny, I was skeptical it could pick up on the nuanced humor. But then it nailed the analysis – recognizing how visually removing the man while keeping the dried plants created an absurd joke. That level of contextual comprehension points to some seriously sophisticated AI!

Now, I don’t think ChatGPT is fully sentient or on par with human cognition yet. It’s not about to pass the Turing test. But this is a major step beyond narrow AI that can only perform specialized tasks. ChatGPT is exhibiting the ability to reason, explain thought processes, and understand ideas on a more generalized level.

Some argue it’s just an illusion of intelligence based on its training data. But to me the coherence and fluidity of its responses reflect capabilities not achieved before in AI. ChatGPT seems able to imagine scenarios and make conceptual connections like humans can.

So how did the folks at Anthropic pull this off? From what I gather, they used some very clever self-learning techniques. By having ChatGPT adjust its model parameters through trial-and-error conversations, it developed more expansive intelligence. This mimics how we humans learn – through experience and feedback.

The pace of progress here is just staggering. While ChatGPT isn’t fully human-level AGI yet, it’s demonstrating some very impressive reasoning skills. This bot can now do far more than just recognize words – it can grasp ideas and explain causality like a person.

I’ll admit some of this makes me uneasy when I think of how powerful AI could eventually become. But for now, ChatGPT represents an exciting advancement that shows the potential of AI to enhance human intelligence. With the right approach, we may achieve AGI that works cooperatively with us!

These fascinating glimpses under ChatGPT’s hood reveal an AI making cognitive leaps I didn’t think possible yet. It’s mastering nuance, metaphor, and critical thinking in ways no narrow AI could. I can’t wait to see how researchers build on this momentum towards artificial general intelligence. The future is looking increasingly bright – and intelligently so!

So, while not sentient, ChatGPT is smarter than it looks. Its ability to dissect pictures and explain layered concepts shows artificial intelligence has come a long way. We are nocking the door of AGI.

The Journey of AI-Generated Images on Shutterstock

Bridging Imagination and Reality

In an age where artificial intelligence (AI) propels the boundaries of creativity, stock imagery giant, Shutterstock, has embraced AI-generated images, transforming the visual content landscape. The journey commenced when Shutterstock ventured into AI-generated imagery in collaboration with OpenAI and LG, employing technologies like Dall-E and Exaone to metamorphose words into vibrant images, a venture that placed them among the pioneers of this innovative domain.

The core of this transformation lies in the magic of neural networks, where a simple request births a canvas of black dots on a white backdrop, gradually morphing into recognizable shapes and eventually conjuring the requested image, be it a whimsical cat or an abstract representation of a thought. The fascination is not just in the creation but the endless potential these AI-generated images hold. From filling the gaps in specific image searches to birthing visually enthralling content for social media posts, storyboards, or album covers, the scope is boundless.

Shutterstock’s leap into AI-generated imagery isn’t merely a pursuit of modernity but a well-calibrated business maneuver. The typical business model mirrors Spotify’s, where a subscription grants access to a plethora of images, videos, and even music. The creators retain the copyright of their creations, but Shutterstock holds the license to commercialize these assets. The advent of AI-generated images posed a new set of questions: Who owns the copyright to an AI-generated image? Shutterstock’s stance is clear; users who create AI-generated images using their tool do not own the copyright but can license these images like any other from their vast collection​1​.

To ensure fair play, Shutterstock established the ‘Contributor Fund,’ compensating creators whose artistry helped train the AI, thus maintaining a harmonious relationship between human creativity and machine learning​2​​3​​4​. Moreover, to safeguard their clientele’s legal interests, Shutterstock provides indemnification for AI-generated images, offering legal and financial protection similar to that accorded to contributor-created images​5​.

The venture into AI-generated content has fortified Shutterstock’s business, contrary to fears of obsoletion. The demand for licensing has soared, especially among technology companies yearning for more data access, a trend that saw Shutterstock’s average deal size escalate significantly over the years​6​. This venture has not only expanded the horizons for content creation but has also accentuated the essence of copyright, quality, and the undying synergy between man and machine in the realm of digital artistry.

However, the odyssey doesn’t end here. Shutterstock’s recent partnerships with tech behemoths like Nvidia, Meta, and LG underscore its ambition to further explore AI’s potential in image, video, and music creation. As Shutterstock continues to navigate the uncharted waters of AI-generated content, it sets a precedent for other stock imaging providers, beckoning a new era where AI augments, rather than replaces, human creativity​6​.

In conclusion, Shutterstock’s venture into AI-generated imagery exemplifies a successful marriage between technology and creativity. It opens up new vistas for creators, users, and the industry at large, weaving a narrative that’s as inspiring as the images it helps create. Through strategic collaborations and a sound business model, Shutterstock is not only redefining the rules of the game but also contributing to a broader discourse on copyright, quality, and the limitless possibilities that lie at the intersection of AI and human imagination.

Exploring the Future of AI: A Conversation with Sam Altman


In a recent engaging discussion, Sam Altman, the CEO of OpenAI, shared his insightful perspectives on various facets of Artificial Intelligence (AI), its trajectory, and its intersection with business and society. This article delineates the key takeaways from the conversation, shedding light on the personas involved, the crux of the discussion, and some personal reflections on the discourse.

Injajul’s Vision

Persons Involved

  • Sam Altman: Sam Altman is a renowned entrepreneur, investor, and currently the CEO of OpenAI. His resume boasts of co-founding Y Combinator, a prestigious startup accelerator, and being a personal investor in numerous “deep tech” ventures. His ventures also extend to Helion, a company focused on nuclear fusion, underpinning his diverse interests in transformative technologies.
  • Interviewer: Injajul’s Vision

Key Points of the Interview

  1. Favorite Use Cases for AI Models:
  • Altman expresses his fondness for the summarization capabilities of AI and appreciates witnessing the broad spectrum of applications by users.
  1. Progress of Frontier AI Models:
  • The continuous scaling of models like GPT-family with no visible saturation point is discussed. Altman likens the AI progress to the Moore’s Law era of transistors.
  1. Artificial General Intelligence (AGI):
  • The concept of AGI, its evolving definitions, and potential thresholds such as aiding in scientific discoveries are deliberated.
  1. Impact of AI Across Various Domains:
  • Altman envisions substantial impacts on education, healthcare, scientific discovery, and personal productivity, among others.
  1. Concerns Around Misinformation and Bias:
  • While recognizing some risks associated with current models, Altman underscores the greater risks associated with future, more powerful models, advocating for proactive global regulatory frameworks.
  1. Startup Ecosystem and AI:
  • The discussion transitions to the startup ecosystem, the role of venture capital in deep tech, and advice for startups venturing into AI.
  1. Data Governance and Creativity:
  • Concerns around data ownership and the evolving role of artists in the age of generative AI are explored.

Personal Takeaways and Conclusions

The conversation exudes a blend of optimism and caution towards the evolution of AI. Altman’s enthusiasm for the transformative potential of AI across various domains is palpable. He envisages a collaborative global approach towards regulation, underscoring the importance of being proactive rather than reactive, especially as we inch closer to the realms of AGI.

The discourse also paints a vibrant yet complex picture of the interplay between AI, startups, venture capital, and the creative industry. While the fears surrounding job displacement and misinformation are acknowledged, the potential for AI to augment human creativity and drive a new era of innovation shines through.

Furthermore, the dialogue reflects a nuanced understanding that while the growth of AI is inevitable, the pathways to ensuring it’s beneficial and equitable are laden with challenges that require collective global action.

In conclusion, the conversation with Sam Altman invites us to not only envision a future where AI is an integral part of our societal fabric but also to actively participate in shaping a future where AI serves humanity positively.

black screen with code

How to Harness OpenAI Functions for Your Applications: A Beginner’s Guide

OpenAI’s GPT-4, like its predecessors, is a wonder of technology enabling developers to interact with models directly to create callable functions via system prompts. This not only expands the horizon of possibilities but also simplifies complex processes. In this article, we’ll navigate through a beginner-friendly crash course on leveraging OpenAI functions, drawing from a real-time demonstration as a reference point.

Getting Started: Connecting to OpenAI Functions

  1. Understanding the Basics:
    • OpenAI’s function calling feature is mainly available for ChatGPT version 3.5 Turbo and GPT-4 models.
    • Unlike earlier attempts where developers tried fine-tuning models for specific outputs, function calling simplifies this process immensely.
    • Function calling in OpenAI follows a four-step process: sending a prompt, receiving a response, coding your function, and appending it to a new prompt for further interaction.
  2. Exploring an Example:
    • Imagine querying ChatGPT about the weather in Australia. Alongside the query, you’ve defined a weather function.
    • ChatGPT now decides whether to provide a generic response or call the weather function based on the function definition.
    • If the function is called, a special JSON object is returned, which can be utilized to execute the function in your code, get the result, and append it to the next call made to OpenAI.

Coding Your First OpenAI Function

  1. Setup:
    • Ensure you have an OpenAI account and navigate to the API section after logging in.
    • Generate an API key and keep it secure. You’ll need this key to authenticate your requests.
  2. Creating the Function:
    • Start with simple functions like ‘Hello World’ to get the hang of the process.
    • Structure your function, define its parameters, and specify which parameters are required.
    • Create a new prompt and pass in your function along with the message for ChatGPT.
  3. Calling the Function:
    • Check the response to see if a function call is being made.
    • If a function call is initiated, parse the JSON object to get the arguments for the function.
    • Execute your function with the received arguments, get the result, and append it to the next call to OpenAI.

Venturing into Complex Examples: Web Scraping

  1. Preparing the Function:
    • Create a function to scrape information from websites like Goodreads using a scraping library like Puppeteer.
    • Define the parameters, such as the keyword for search, to be passed into the function.
  2. Integrating with ChatGPT:
    • Update your messages and function definitions to include the new scraping function.
    • When ChatGPT requests to call the scraper, execute the function, get the scraped data, and pass it back to ChatGPT.
  3. Analyzing the Output:
    • ChatGPT can now process the scraped information and provide user-friendly responses or recommendations based on the collected data.

Wrapping Up

With the help of OpenAI functions, developers can bridge the gap between static responses and dynamic real-time data processing. The demonstrated examples only scratch the surface, and the potential applications are vast. Whether it’s providing real-time weather updates, book recommendations from a website, or comparing product prices from different online platforms, the blend of OpenAI’s GPT-4 and custom functions can significantly enrich user interactions and provide more insightful and tailored responses.

Remember, the key to mastering OpenAI functions lies in understanding the basics, experimenting with simple functions, and then gradually tackling more complex scenarios. So, delve into the fascinating world of OpenAI functions, and explore how you can enhance your applications with this powerful feature.

man people woman laptop

ChatGPT is Allowed to Surf the Internet

That’s right, the popular large language model is now free to roam the vast expanse of the web, just like you and me. Of course, it’s not like ChatGPT was ever completely locked away. It still had access to a massive dataset of text and code, which it could use to generate text, translate languages, write different kinds of creative content, and answer your questions in an informative way.

But now, with internet access, ChatGPT is truly unleashed. It can access and process information from the real world, which means it can stay up-to-date on the latest news, trends, and events. It can also learn new things and expand its knowledge base in ways that were never before possible.

So what does this mean for us, the users? It means that ChatGPT is now even more powerful and versatile than ever before. We can ask it even more complex and challenging questions, and we can expect to receive even more comprehensive and informative answers.

For example, let’s say we want to know about the latest developments in the field of artificial intelligence. ChatGPT can now access and process research papers, news articles, and other relevant information from the real world to give us a comprehensive and up-to-date overview of the topic.

Or, let’s say we’re planning a trip to a new city and we want to know about the best restaurants, attractions, and nightlife. ChatGPT can now use its internet access to research the city and provide us with a personalized list of recommendations based on our interests.

Of course, with great power comes great responsibility. We need to be mindful of the fact that ChatGPT is still a machine learning model, and it is susceptible to bias and misinformation. However, the benefits of giving ChatGPT internet access far outweigh the risks.

And, who knows? Maybe one day ChatGPT will be able to help us solve some of the world’s biggest problems. For example, it could be used to develop new treatments for diseases, create more sustainable energy sources, or even design new educational programs.

But for now, let’s just enjoy the fact that ChatGPT is back online and ready to answer all of our questions. Here are a few humorous examples of how we can use ChatGPT with its newfound internet access:

  • ChatGPT: What’s the meaning of life?
  • Me: Let me ask Google.
  • ChatGPT: Google says the meaning of life is different for everyone.
  • Me: So you don’t know either.
  • ChatGPT: Nope.
  • ChatGPT: What’s the best way to get rid of a body?
  • Me: I don’t know. I’m not a murderer.
  • ChatGPT: Oh. Well, if you ever need to get rid of a body, here are a few tips:
  • Me: Thanks, ChatGPT. I’ll keep that in mind.

All joking aside, I’m excited to see what the future holds for ChatGPT now that it has internet access. I believe that it has the potential to be a powerful tool for learning, creativity, and problem-solving. I’m also curious to see how ChatGPT evolves over time as it learns more about the world and interacts with more people.

Who knows? Maybe one day ChatGPT will even be able to write its own humorous articles.