ChatGPT’s Rival Now Available Free of Charge…

🤖 AI tools

Edit Template

ChatGPT's Rival Now Available Free of Charge

ChatGPT's Rival Now Available Free of Charge

In today’s newsletter, we’ll explore how to create websites in a matter of seconds and how preventing bank fraud before it happens is no longer a pipe dream. This is the kind of artificial intelligence we like, the kind that leads us towards a better future.

What am I talking about? Let’s find out together.

🗞️ The news

1/

Wix and all the new features.. Obviously, powered by AI.

In a new blog post, Wix CEO Avishai Abrahami outlined what we can consider a comprehensive overview of the platform’s features developed through the use of AI.

Among those already available are text and image creation, background removal, and product descriptions, extending to video editing.

However, what truly struck me are two new features under development. The creation of complete websites through a prompt and a personal business assistant.

The first one will allow you to create everything (texts, templates, images, sections) through a simple prompt, as you can see in this video. We’re not talking about predefined templates, but something tailor-made for the user.

The second one, on the other hand, will suggest improvements for the website and create strategies based on personalized analyses and site trends.

2/

Mastercard’s new “anti-fraud” system could prevent millions of losses for banks

The initiative is called “Consumer Fraud Risk” and aims to help UK banks predict and prevent scams in real-time payments.

The system overlays payment data with analysis variables such as account names and payer and recipient histories.

This helps banks intervene to stop fraudulent transactions before the money is lost.

In particular, the system has the potential to reduce the number of fraud cases where a scammer impersonates a friend, relative, or company and induces the victim to transfer money to them.

These cases currently account for 40% of bank fraud losses in the UK.

3/

Meta releases Llama 2 and strengthens its partnership with Microsoft

A week that could change the entire world of artificial intelligence. In fact, Meta announced that its language model Llama 2, considered one of the direct competitors of ChatGPT, will be available both on the Microsoft Azure platform and as an open-source model downloadable by anyone.

Not only that, but Meta has also authorized its commercial use, provided that the product does not exceed 700 million monthly active users (MAU).

To give you an idea, Snapchat has reached 750 MAU, and TikTok is projected to reach about 800 million MAU, both just this year.

🤑 Funding of the week

Preply secures $120M in funding

Preply is a language learning app known for its live tutors, which has seen its sales grow tenfold just in the last three years.

The company, renowned for linking students to a network of roughly 35,000 tutors and group classes across nearly 50 languages, intends to use the funds to further reinforce its dedication to AI

“While human-led learning will always be at Preply’s core, the company is using AI as an engine to scale learning, improve access and focus on teaching like never before

🧠 AI World News

🧪 Prompt corner

This week we continue with ChatGPT’s code interpreter, but not with data analysis.

We will indeed see how to turn images into a slideshow video.

What we need to do is gather all our images into a single folder and compress it into a .zip file (enabling us to upload it as a singular file).

We then upload it using the following prompt:

This week we continue with ChatGPT’s code interpreter, but not with data analysis.

We will indeed see how to turn images into a slideshow video.

What we need to do is gather all our images into a single folder and compress it into a .zip file (enabling us to upload it as a singular file).

We then upload it using the following prompt:

📸 The Image of the Week

Hip Hop and R&B Stars Reimagined as Lego Characters 🤖

⚒️ Useful Resources

One place for all the tools you need. (link)

A course from 0 to 1 on prompting with ChatGPT (link)

Learn prompting from A to Z (link)

How to use Midjourney to generate images (link)

SnackPrompt – Discover the best prompts for Chat-GPT (link)

Improve your life with Artificial Intelligence.

Every Thursday, your weekly dose of Artificial Intelligence. Receive only the most interesting information, news and tools about the world of AI.

© AI onboard – Privacy PolicyCookie Policy

VAT 04309790402

AI Apocalypse, can we avoid It this way?

AI Apocalypse, can we avoid It this way?

AI Apocalypse, can we avoid It this way?

Just when we find ourselves talking more and more about OpenAI’s competitors, THE novelty arrives: the code interpreter. But that’s not all…

🗞️ The news

1/

Now anyone can be a data analyst thanks to the Code Interpreter.

The new feature has been made available in Beta Test for all Plus users of the platform.

It will allow you to create and run Python scripts, upload files directly to ChatGPT, analyze data, create code and modify it manually (or by asking to do so) directly on the platform.

The possibilities with this new feature, as you can imagine, are endless.

Since its release, all I’ve been doing is analyzing data, almost spasmodically, to then create graphic visualizations that clarify the results.

In the prompt corner we’ll see together how to do it.

Do you think it could be useful to you in some way? Reply to this email and tell me how!

2/

The other side of the coin is called Superalignment.

On one hand, there’s excitement for the code interpreter, on the other, the creation of a team specialized in controlling the immeasurable expansion of this technology.

In fact, according to OpenAI, superintelligence, i.e., AI with intelligence superior to human, could arrive in the next decade.

The team, led by Ilya Sutskever and Jan Leike, will have 20% of the company’s computing power and 4 years to figure out how to control these AIs.

The idea is to use human feedback to train the artificial intelligence, use this AI to supervise other AI systems and develop another AI dedicated to detecting problematic behavior or internal inconsistencies.

Although not certain of achieving their goal, OpenAI is optimistic and believes that a focused and concentrated effort can solve the problem.

3/

Volkswagen has begun testing autonomous driving of its vehicles in Austin

Remember when in Back to the Future Marty McFly arrives in the “futuristic” 2015 and sees flying or autonomously driven cars? Well, they may have gotten the decade wrong.

In fact, Volkswagen has started a test program for autonomous driving of its ID Buzz model in the city of Austin.

The test began this month with 2 vehicles and is expected to reach 10 by the end of the year.

The vehicles, equipped with cameras, lidar, radar, and autonomous driving software, will initially be used to collect data and will always have a human operator behind the wheel for safety reasons.

Source: Volkswagen

🤑 Funding of the week

Telexistence Secures $170M in Funding

The Japanese startup specializing in robotics has secured funding of $170 million, bringing the total funds raised to about $193.5M since the start of its operations.

Supporters include SoftBank, Airbus Ventures, Monoful Partners, KDDI Open Innovation Fund (a fund created by Foxconn and CTBC Financial Holdings), and Globis Capital Partners.

The company will use the raised funds to strengthen its global workforce, as CEO Jin Tomioka has stated.

In addition to the funding, Telexistence has announced a strategic partnership with SoftBank Robotics Group, a division of SoftBank Group, to accelerate its commercialization in North America and Foxconn to produce its next-generation robot, named “Ghost”.

🧠 AI World News

🧪 Prompt corner

There’s no escape. This week we’re talking about the code interpreter and how to use it to analyze data and aid us in understanding them through graphical formatting.

What I’ll show you is the “how,” but the “what” to analyze has no limits. Let’s get started.

Suppose we want to analyze Amazon’s shares over the past 5 years and want a chart that clearly shows us the trend of value over time.

So we will look for a dataset that contains this information. For today’s example, I found an old dataset here: https://www.kaggle.com/datasets/varpit94/amazon-stock-data

As you can see, I asked to analyze the last 5 years. So ChatGPT will independently filter only the desired period.

Once we get to this stage, we could continue the analysis, asking for more information or integrating other data such as news, acquisitions, new product presentations, and so on.

📸 The Image of the Week

What if the story of Shrek were “based on a true story”?

⚒️ Useful Resources

One place for all the tools you need. (link)

A course from 0 to 1 on prompting with ChatGPT (link)

Learn prompting from A to Z (link)

How to use Midjourney to generate images (link)

SnackPrompt – Discover the best prompts for Chat-GPT (link)

Improve your life with Artificial Intelligence.

Every Thursday, your weekly dose of Artificial Intelligence. Receive only the most interesting information, news and tools about the world of AI.

© AI onboard – Privacy PolicyCookie Policy

VAT 04309790402

Gone in 5 years…

Gone in 5 years...

Gone in 5 years...

We all know that artificial intelligence is revolutionizing the world. However, the matter becomes especially serious when the CEO of Stability AI, one of the most established companies in the sector, predicts that the role of developers could become obsolete within five years. But let’s look at it together…

🗞️ The news

1/

How AI will change the world of work (and not only) in the near future

In an interview with Peter H. Diamandis, the CEO of Stability AI shared his point of view on the future of artificial intelligence.

AI is revolutionizing the world in a way never seen before, and for some jobs, the end may have come.

In fact, today about 41% of code is generated by AI, and according to Emad, “in 5 years, there will be no more programmers”.

Not only that, but he also predicts that within a year, we will have a form of artificial intelligence similar to ChatGPT in our phones, working even without an internet connection.

2/

Chat GPT’s web browsing has been disabled, it freely showed paid content.

On July 4th, OpenAI announced that it had disabled its web browsing function due to some reports received in recent days.

Some users had reported the possibility of bypassing the paywalls of various news outlets simply by asking ChatGPT to print the content of the page.

If web browsing is a function you often used, don’t worry, in today’s prompt corner I will explain how to continue using it.

 

3/

How Meta uses AI to influence what you see on Facebook and Instagram

In the article published by Nick Gregg, it is explained how AI is used to make the app experience “unique and personalized”. In particular, I found the following points interesting:

  • Meta is working to be more transparent about the AI systems that incorporate user feedback to classify content on Facebook and Instagram.
  • The AI systems predict how valuable a piece of content may be to a user so that it can be displayed first.
  • Meta will expand the “Why am I seeing this?” function in the coming weeks on Instagram Reels, Explore and Facebook Reels, after previously launching it for some Feed content and all ads.

🤑 Funding of the week

 

Inflection AI secures $1.3B funding

Inflection AI, an artificial intelligence startup led by Mustafa Suleyman, former DeepMind leader, has secured funding of $1.3 billion, bringing its valuation to $4 billion.

Supporters include names such as Microsoft, Nvidia, and tech moguls Reid Hoffman, Bill Gates, and Eric Schmidt.

The Palo Alto-based company launched its first chatbot, Pi, less than two months ago and will use the funds raised to enhance its computing capabilities and further develop it.

Inflection is collaborating with Microsoft, its cloud computing partner, and Nvidia, which has assisted Inflection in implementing its H100 graphic processing unit, considered the gold standard for AI training.

With the help of Nvidia and CoreWeave, Inflection plans to install thousands of additional GPUs, and once completed, Inflection’s new cluster, consisting of 22,000 H100s, will be the largest in the world.

🧠 AI World News

🧪 Prompt corner

As anticipated in the news section, this week’s prompt corner will be short but hopefully useful to many.

True, web browsing has been temporarily disabled, but the plugins are still there, ready to be used.

In fact, the ones we’ll see today are the best plugins in the entire store for online browsing.

There are 2 plugins with similar functions, in order of personal preference:

BrowserPilot and WebPilot

📸 The Image of the Week

What would today’s VIPs look like if they had lived in the Neanderthal era?

⚒️ Useful Resources

One place for all the tools you need. (link)

A course from 0 to 1 on prompting with ChatGPT (link)

Learn prompting from A to Z (link)

How to use Midjourney to generate images (link)

SnackPrompt – Discover the best prompts for Chat-GPT (link)

Improve your life with Artificial Intelligence.

Every Thursday, your weekly dose of Artificial Intelligence. Receive only the most interesting information, news and tools about the world of AI.

© AI onboard – Privacy PolicyCookie Policy

VAT 04309790402

Has the anti-ChatGPT-4 arrived? It’s called Gemini..

Has the anti-ChatGPT-4 arrived? It's called Gemini..

Has the anti-ChatGPT-4 arrived? It's called Gemini..

If I had to bet on which company would create the real competitor to GPT-4, my money would undoubtedly be on Google. Will Gemini mark the end of OpenAI?

🗞️ The news

1/

Google Deepmind’s CEO announces that their chatbot will be better than ChatGPT-4

The artificial intelligence company announced that they are working on an AI capable of performing better than OpenAI’s chatbot.

It will be called Gemini and has required tens of millions of dollars for its development, a cost comparable to ChatGPT.

Demis Hassabis, CEO of the company, emphasized how this AI benefits from the extensive past experience gained in the field of artificial intelligence.

In particular, Gemini uses techniques developed for AlphaGo, a program known for its performance in the board game Go.

The main technique used in AlphaGo and now implemented in Gemini is the “reinforcement learning”, in which repeated attempts are made to improve performance through a system of positive and negative rewards.

 

2/

Adobe takes legal responsibility for all content generated through Firefly’s artificial intelligence

Especially when companies use this technology, the fear that the content could lead to legal issues is real.

Adobe has understood this and announced that With Firefly, Adobe will also be offering enterprise customers an IP indemnity, which means that Adobe would protect customers from third party IP claims about Firefly-generated outputs.”

How does this translate into practice? From the user’s side (the ones with the enterprise plan), you can proceed with the creation of any content while sleeping soundly, while from Adobe’s side, this move certainly conveys a sense of transparency regarding the database used to train its AI model and consequently inevitably attract new customers.

 

3/

The main AI models do not respect some of the laws of the new draft of the AI ACT

After the announcement a few weeks ago, the first draft of the AI ACT arrived. Stanford University therefore decided to conduct a study to assess the state of the art of the world’s largest AI companies.

The most popular models like GPT-4, PaLM 2, and Claude only met a measly 50% of the necessary criteria.

But why is this so important? In addition to guaranteeing privacy, there’s a risk of a $20M (or 4% of total global turnover) fine.

🤑 Funding of the week

MosaicML is acquired by Databricks for $1.3 billion

This week more than a financing, we will see a real acquisition.

MosaicML is a company that offers a platform to build, train, and implement AI models based on proprietary data.

It also offers a series of “preset” open-source models that can be easily completed and used.

All these services will therefore be incorporated under the “Databricks” umbrella, which in turn offers technology for the storage of all company data.

This will allow the creation of a more complete service. The users will have a place to store their data and at the same time being able to use them to train artificial intelligence.

🧠 AI World News

🧪 Prompt corner

Each of us sells something every week, be it a product or our expertise.

Creating content on social media is crucial for this endeavor.

This week, we’re going to ask GPT-4 to write a script for our videos:

I’d like you to act as an experienced social media manager. I’m / I own a [Insert product/job] and would like you to give me 5 viral ideas for Instagram Reels or TikTok, and write a script for each of them.

Each script must include:

  1. An attention-grabbing hook
  2. The main body
  3. A final call to action encouraging people to follow me

📸 The Image of the Week

“New York City nightlife in the year 2090”

⚒️ Risorse utili

One place for all the tools you need. (link)

A course from 0 to 1 on prompting with ChatGPT (link)

Learn prompting from A to Z (link)

How to use Midjourney to generate images (link)

SnackPrompt – Discover the best prompts for Chat-GPT (link)

Improve your life with Artificial Intelligence.

Every Thursday, your weekly dose of Artificial Intelligence. Receive only the most interesting information, news and tools about the world of AI.

© AI onboard – Privacy PolicyCookie Policy

VAT 04309790402

Too Dangerous to Make Public…

Too Dangerous to Make Public...

Too Dangerous to Make Public...

Creating an artificial intelligence model and then realizing that it’s too dangerous to make public is not something that happens often at Meta, yet that’s exactly what happened this week. The other social media giant TikTok, on the other hand, just ordered… 100,000 GPUs from Nvidia to get a foothold in the big AI game.

🗞️ The news

1/

Meta presents a new text-to-speech AI model, but it won’t be made public due to its dangerousness.

Last week, Meta introduced Voicebox, an artificial intelligence tool that allows translating any textual input into vocal output.

However, as explained in a blog post, they stated that “there are many possibilities of use, but due to the potential risks in case of misuse, we have decided not to make the Voicebox model or its code public, at this time.”

In fact, the AI is considered so technologically advanced that it could enable the easy spread of fake news or the execution of online scam

2/

 

Mercedes introduces ChatGPT-based artificial intelligence in its cars.

Thanks to a collaboration with Microsoft, the goal is to have a voice assistant that is not just a database of predefined questions and answers.

The aim is to improve the understanding of “natural” language and widen the sphere of topics to which the assistant can respond. In short, you can chat about traffic and last night’s soccer match.

In America, the beta phase for all Mercedes owners with MBUX infotainment has already started a few days ago. To enable the installation, a simple “Hey Mercedes, I want to participate in the beta program” will be sufficient.

3/

Chinese giants, including TikTok and Alibaba, are buying more than $1 billion in NVIDIA chips.

In the first 6 months of 2023, ByteDance (the company that owns TikTok) alone ordered about 1 billion dollars of Nvidia GPUs, equivalent to about 100,000 units.

This value alone corresponds to the entire Chinese market for Nvidia in 2022.

The orders are divided between the A100s, ordered before the US restrictions on high-performance GPU exports, and the H800s, a model created specifically to “respect” them.

Hardware investments are motivated by the desire to compete with the West in the field of artificial intelligence, both to take a slice of the $4.4 trillion a year that AI could bring to the economy (we talked about it in the last newsletter) but also because China needs high-tech solutions for its state control apparatus, such as CCTV analysis and facial recognition technology.

🤑 Funding of the week

We continue to talk about dizzying numbers, this time it’s the turn of Mistral AI, a French company that after only 4 weeks of life receives $113M in funding and brings its value to $260M.

Founded by former employees of Deep Mind and Meta, its goal is to compete with giants like OpenAI in creating and implementing large-scale linguistic models and generative AI, especially for businesses.

The biggest challenge? According to CEO Arthur Mensch, simply “making artificial intelligence useful”. To do this, they will release their first AI models in 202​4.

🧠 Miscellaneous from the world of AI.

🧪 Prompt corner

This week I would like to do an experiment. Below you will find the tutorial to create Qr codes like this.

The whole thing is created through

Finished browsing
ControlNet, a model usable in Stable Diffusion.

Compared to previous weeks, it’s technically more complicated, which is why today more than ever it’s important that you give me feedback on the level of the prompt by writing to me at newsletter@aionboard.com

📸 Picture of the week

Le città europee, disegnate dall’intelligenza artificiale 🖌️

⚒️ Risorse utili

A single place, all the tools you need (link)

A course from 0 to 1 on prompting with ChatGPT (link)

Learn prompting from A to Z (link)

How to use Midjourney to generate images (link)

SnackPrompt – discover the best prompts for Chat-GPT (link)

How to use notion AI to have more time in your days​ (link)

Improve your life with Artificial Intelligence.

Every Thursday, your weekly dose of Artificial Intelligence. Receive only the most interesting information, news and tools about the world of AI.

© AI onboard – Privacy PolicyCookie Policy

VAT 04309790402

© AI miei tempi – Privacy PolicyCookie Policy

P.IVA 04309790402