Many companies are experimenting with deploying AI and more and more examples of applications are coming out.
A month ago, for example, Klarna Bank launched its AI assistant, powered by OpenAI, worldwide. With already impressive results:
The AI assistant has made 2.3 million calls (this is two-thirds of all chats handled by Klarna's customer service).
The assistant does the work of 700 full time employees.
In terms of customer satisfaction score, it is equivalent to human employees.
It is more accurate in solving problems, leading to 25% fewer repeated queries.
Customers now solve their problems in less than 2 minutes, compared with 11 minutes previously.
The assistant is available 24/7 in 23 countries and communicates in more than 35 languages.
It is estimated that this will generate a $40 million profit improvement for Klarna by 2024!
Klarna has integrated this AI assistant into their app to improve the user experience for their 150 million consumers worldwide. The assistant offers a wide range of services, including multilingual support, handling refunds and returns, and encouraging responsible financial behavior.
This introduction is seen by Klarna as a major step forward in their vision of a fully AI-powered financial assistant designed to save consumers time, worry and money, while making the global banking industry more efficient and consumer-focused.
Sebastian Siemiatkowski, co-founder and CEO of Klarna, said:
"This breakthrough in AI for customer interaction means superior experiences for our customers at better prices, more interesting challenges for our employees and better returns for our investors. We are incredibly excited about this launch, but it also underscores the profound impact AI will have on society.
We want to emphasize and encourage society and politicians to consider this carefully. Thoughtful, informed and stable management will be crucial to navigate through this transformation of our societies."
In an interview early last year, Sam Altman, CEO of OpenAI, already expressed the expectation that customer service is one of the areas in which there will be far fewer jobs in the near future. An example such as Klarna's AI assistant underscores this expectation, and numerous other examples will follow this year.
Take a leap forward in your marketing AI transformation every week
Every Friday, we bring you the latest insights, news and real-world examples on the impact of AI in the marketing world. Whether you want to improve your marketing efficiency, increase customer engagement, sharpen your marketing strategy or digitally transform your business, "Marketing AI Friday" is your weekly guide.
https://nxtli.com/wp-content/uploads/2024/03/Klarna-AI-assistent.jpeg450800Michiel Schoonhovenhttps://nxtli.com/wp-content/uploads/2019/10/NXTLI_Next_Level_Impact_Logo-1-300x123.pngMichiel Schoonhoven2024-03-01 20:01:022024-05-10 16:02:05Klarna AI assistant replaces human customer service
Today OpenAI introduced Sora, their impressive text-to-video model. With Sora, you can very easily generate videos up to a minute long based on a text instruction you provide as a user.
See below an example of a video created from the text instruction:
"A litter of golden retriever puppies playing in the snow. Their heads pop out of the snow, covered in."
Or how about this one:
'Beautiful, snowy Tokyo city is bustling. The camera moves through the bustling city street, following several people enjoying the beautiful snowy weather and shopping at nearby stalls. Gorgeous sakura petals are flying through the wind along with snowflakes.'
Sora is able to generate complex scenes with multiple characters, specific types of movements and accurate details of the subject and background. The model understands not only what the user has requested in the prompt, but also how those things exist in the physical world.
The model has a deep understanding of language, enabling it to accurately interpret prompts and generate compelling characters that express vivid emotions. Sora can also create multiple shots within a single generated video that accurately preserve characters and visual style.
Sora was only made available to a select group of users yesterday, but is being rolled out incrementally.
Take a leap forward in your marketing AI transformation every week
Every Friday, we bring you the latest insights, news and real-world examples on the impact of AI in the marketing world. Whether you want to improve your marketing efficiency, increase customer engagement, sharpen your marketing strategy or digitally transform your business, "Marketing AI Friday" is your weekly guide.
https://nxtli.com/wp-content/uploads/2024/02/Schermafbeelding-2024-02-15-om-21.49.22.png11622088Michiel Schoonhovenhttps://nxtli.com/wp-content/uploads/2019/10/NXTLI_Next_Level_Impact_Logo-1-300x123.pngMichiel Schoonhoven2024-02-15 20:06:002024-05-10 16:02:14OpenAI introduces text-to-video model Sora
Recent research from BrightEdge shows that Google's Search Generative Experience (SGE) is going to fundamentally change the way we search, a "Searchquake" in the digital world! It is critical that companies prepare for the impact SGE will have on brand visibility and customer behavior.
SGE allows you to perform searches more efficiently by instantly providing comprehensive answers, reducing the need for multiple searches and clicking through to websites. In this regard, AI acts as a bridge between your content and the Google user.
For example, we are going to see that AI can simulate a complete shopping experience, guiding visitors step by step without requiring them to leave the search engine.
The study shows that 84% of Google searches will be affected by SGE and that the healthcare, e-commerce and B2B technology sectors will be most affected.
SGE makes the visibility of Web sites in search results more complex, which can lead to a decrease in the number of visitors to your Web site through search engines. However, the visitors you still attract may be more valuable and engaged.
The shift to SGE is not just Google's next algorithmic change. It is the most significant search engine change ever and represents a profound change that will affect every industry, every business and every marketer.
It is essential that you adjust your SEO strategy and website experience accordingly. To prepare for SGE, here are some action items:
Expertise: Show that the content on your website was created by people with in-depth knowledge of the topic. For example, share extensive biographies of authors and relevant qualifications.
Experience: Share evidence of practical and relevant experience, such as case studies or customer stories.
Authoritativeness: Gain authority through recognition from leading websites and achieving certain certifications or standards.
Trustworthiness: Secure your website, be transparent in communication and collect positive reviews.
Optimize for conversational search and long-tail keywords:
Use natural colloquialisms in your content.
Anticipate literal questions from your audience and integrate them into your content.
Write in the second person to create a personal dialogue.
Enrich your content with storytelling and interactive elements such as chatbots.
Be clear and concise in your message.
Integrate multimedia content:
Use videos, podcasts and infographics to make answers more lively.
Analyze search intent:
Track how search intentions change and adjust your content strategy accordingly.
Build your online reputation:
Collect customer reviews and mentions in leading publications.
Provide contextual relevance:
Create in-depth and informative content that not only answers questions, but also understands the broader context.
Stay flexible:
Be prepared to adapt your strategies to the rapidly evolving digital marketing world.
By applying these recommendations, you will ensure that your SEO strategy and website experience are ready for the future of search with SGE, and provide a more valuable experience to your visitors.
Take a leap forward in your marketing AI transformation every week
Every Friday, we bring you the latest insights, news and real-world examples on the impact of AI in the marketing world. Whether you want to improve your marketing efficiency, increase customer engagement, sharpen your marketing strategy or digitally transform your business, "Marketing AI Friday" is your weekly guide.
https://nxtli.com/wp-content/uploads/2024/02/Searchquake.webp458462Michiel Schoonhovenhttps://nxtli.com/wp-content/uploads/2019/10/NXTLI_Next_Level_Impact_Logo-1-300x123.pngMichiel Schoonhoven2024-02-12 07:21:002024-05-10 16:02:21Prepare for a 'Searchquake'
Google yesterday launched their most advanced Ultra language model under the name Gemini Advanced. In addition, they changed the name Bard to Gemini.
What is Gemini Advanced and what are the initial experiences?
Gemini Advanced
Gemini Advanced provides access to Google's most powerful AI model: Ultra 1.0. With this, it offers many more capabilities on highly complex tasks including:
Programming
Logical reasoning
Following nuanced instructions
Creative collaboration
In addition, Gemini Advanced offers the ability to upload files, documents, data and more for deeper analysis. And it integpricing with Google apps such as Gmail, Maps and YouTube. You can communicate with Gemini via text, your voice or images.
Google indicates that Gemini Advanced will continue to expand in the coming months with new and exclusive features, such as increased context length, expanded multimodal capabilities, and even better programming features.
A Gemini app has also been launched, however it is not yet available in the Netherlands. The browser version is available in the Netherlands though, which you can use to test Gemini Advanced right now. However, it seems that we cannot use the optimized version at this time. Google also indicates that Ultra 1.0 is only available and optimized for English, but can answer in the other languages such as Dutch.
My initial test results showed that Gemini Advanced is not currently equivalent to ChatGPT4.
The first test was a question about current events.
Here Gemini failed completely, while ChatGPT-4 gave an accurate answer.
During testing, I was sitting on the couch and took a picture of 2 tennis balls lying next to me and put there by our dog. To use this to test recognizing and giving context to an image. See below the response of Gemini Advanced versus GPT4.
n this example, a clear difference can also be seen between Gemini Advanced and ChatGPT-4, where the latter (except for text recognition), recognizes the context of the image very well and Gemini has very limited success in this. 🤨
A follow-up question to test the contextual power of both tools was "how do you think those balls got on the bench?
Again, ChatGPT-4 gives a very correct answer, where Gemini's answer is very general.
A subsequent test with an image gave similar results and clearly shows the current quality difference between Gemini Advanced and ChatGPT-4.
Fortunately, Gemini itself also states "I'm still in the midst of development, so it's possible I haven't seen everything shown in the image." but it's clear that Gemini still has quite a few strokes to make, which I'm sure will happen in the near future.
In any case, it is definitely worthwhile to go test Gemini and experiment what it can already do and repeat this regularly. You can test Gemini Advanced for free for 2 months now, after that the cost is Euro 21.99 per month per user.
Take a leap forward in your marketing AI transformation every week
Every Friday, we bring you the latest insights, news and real-world examples on the impact of AI in the marketing world. Whether you want to improve your marketing efficiency, increase customer engagement, sharpen your marketing strategy or digitally transform your business, "Marketing AI Friday" is your weekly guide.
The new year has barely begun and the beacons in the AI landscape are already being moved. OpenAI introduced two interesting developments last Wednesday: the ChatGPT Team subscription as well as the GPT Store. These extensions are an important step in accelerating the adoption of advanced AI technologies.
Let's look at how these innovations can transform the way we work and communicate.
ChatGPT Team
The ChatGPT Team subscription can accommodate up to 149 users, which means you can now easily use a shared ChatGPT environment with your entire team.
You can manage users and team data and chats remain private as they are not used for language model training.
All the benefits of ChatGPT Plus are included, such as access to DALL-E 3, GPT-4 with Vision, and advanced data analysis, but with a significantly larger "context window" of 32,000 tokens. This allows better performance of complex tasks involving extensive conversations or documents.
In addition, it allows team members to develop and share custom GPTs, enabling a whole new form of collaboration and innovation.
You can choose a flexible monthly subscription of $30 per user (starting from 2 users) or a yearly subscription of $25 per user per month.
Existing ChatGPT users can easily upgrade by logging into their account.
Late last year, OpenAI already introduced the ability to develop your own custom GPTs, such as a "DenisDoelandGPT," for example.
Now individual developers and organizations can offer their own customized GPT models in the GPT Store. This makes them accessible to other users who can thus find specific models to suit their needs.
GPT stands for "Generative Pre-trained Transformer" and is simply put, a kind of language model for artificial intelligence to generate answers to questions.
Custom GPTs allow you to create your own versions of ChatGPT "without any coding" that combine instructions, additional knowledge and skills.
These GPTs can help you in your daily activities, in specific tasks, for yourself and for your business.
Again, there is a lot to discover and experiment with.
Take a leap forward in your marketing AI transformation every week
Every Friday, we bring you the latest insights, news and real-world examples on the impact of AI in the marketing world. Whether you want to improve your marketing efficiency, increase customer engagement, sharpen your marketing strategy or digitally transform your business, "Marketing AI Friday" is your weekly guide.
https://nxtli.com/wp-content/uploads/2024/01/OpenAI-GPT-Store.jpg607800Michiel Schoonhovenhttps://nxtli.com/wp-content/uploads/2019/10/NXTLI_Next_Level_Impact_Logo-1-300x123.pngMichiel Schoonhoven2024-01-12 13:02:572024-05-10 15:59:28OpenAI launches ChatGPT Team and GPT Store
This year, artificial intelligence (AI) is going to be increasingly integrated into the systems you use every day. Think Microsoft applications like Word, Excel, Powerpoint and Outlook, and Google applications like Docs, Sheets and Gmail.
Search engines such as Google Search will change dramatically with the integration of AI. In addition, the latest laptops and smartphones will also be equipped with AI technology.
Microsoft, for example, has already announced that their new Surface laptop will be capable of "next-gen AI. Apple says it is working on advanced AI language models for iPhones and other devices. And Samsung is about to unveil the first AI-powered smartphones on Jan. 17 with their latest Galaxy S24 line.
OpenAI announced yesterday that starting next week, it will launch the GPT Store, where individual developers and organizations can offer their own customized GPT models. This increases the accessibility and diversity of AI tools, allowing users to find specific models to suit their unique needs and projects.
In short, it promises to be a year in which AI takes a big leap forward in adoption.
How are you preparing for this AI-driven future?
Take a leap forward in your marketing AI transformation every week
Every Friday, we bring you the latest insights, news and real-world examples on the impact of AI in the marketing world. Whether you want to improve your marketing efficiency, increase customer engagement, sharpen your marketing strategy or digitally transform your business, "Marketing AI Friday" is your weekly guide.
https://nxtli.com/wp-content/uploads/2024/01/AI-integraal-onderdeel-meeste-systemen-die-je-gebruikt.png10241024Michiel Schoonhovenhttps://nxtli.com/wp-content/uploads/2019/10/NXTLI_Next_Level_Impact_Logo-1-300x123.pngMichiel Schoonhoven2024-01-05 16:35:022024-05-10 15:59:28AI becomes an integral part of most systems you use
The expression "garbage in, garbage out" also applies to the use of AI. If you give AI tools incomplete or unclear input, you cannot expect high-quality output. To improve the quality of AI results, it is important to give clear and precise instructions. Sometimes you have to treat AI like a child to whom you have to explain something step by step so that it understands exactly what you mean.
You can apply the 6 techniques below to give better instructions, which will lead to better AI results.
1. Give clear directions
Language models cannot guess what you are thinking. If you want short answers, ask for these. If you want more expert answers, ask for expert-level explanations. If you want a certain type of answer, show what it should look like. The clearer you are, the better the answer you will get.
Tips:
Tell exactly what you want to know to get better answers.
Ask the model to impersonate a particular character.
Use punctuation to make different parts of your question clear.
Describe the steps required for what you want to do.
Give examples.
Say how long or short you want the answer to be.
Below you can see some examples.
Too limited
Better
How do I add numbers in Excel?
How do I add up a row of euro amounts in Excel? I want to do this automatically for an entire sheet of rows, with all totals ending on the right side in a column called "Total."
Who is the president?
Who was Mexico's president in 2021, and how often are elections held?
Write code to calculate the Fibonacci sequence.
Write a TypeScript function to efficiently compute the Fibonacci sequence. Comment the code in detail to explain what each piece does and why it is written that way.
Summarize meeting minutes.
Summarize the minutes of the discussion in one paragraph. Then, for each speaker, describe their main points. Finally, list the next steps or possible action items suggested by the speakers.
2. Use sample texts
Language models can sometimes give wrong answers, especially on difficult topics or when they have to give quotes or Web addresses. Just as cheat sheets can help a student on a test, giving these models a text can help them give better answers without making things up.
Tips:
Have the model answer with information from a given sample text.
Have the model respond with quotes from the sample text.
3. Make difficult tasks easier by breaking them into smaller pieces
Just as programmers break down complicated computer programs into smaller, easier-to-manage pieces, you can do the same with tasks for a language model. Difficult tasks often go wrong and are more difficult than easy tasks. Also, you can often divide difficult tasks into a series of easier tasks. Here you use the outcome of one task to make the next task easier.
Tips:
Determine what is important in a user's question to find the appropriate instructions.
For long chat conversations, list or choose the most important parts of previous conversations.
Make step-by-step summaries of long texts and merge them into a complete summary.
4. Give the model time to think
If you need to quickly calculate how much 17 times 28 is, you may not know the answer right away, but you can figure it out if you take a moment. Similarly, language models make more mistakes if they have to answer immediately. By putting the model through a "thinking process" first, it arrives at better and more reliable answers.
Tips:
Let the model come up with its own solution first, rather than immediately giving an answer.
Have the model think "out loud" or ask a few questions to understand how it arrives at an answer.
Ask the model if it has overlooked anything in previous attempts.
5. Use external tools
Improve the performance of the language model by combining it with other tools. For example, a text retrieval tool can give the model information about relevant documents. A program to execute code, such as OpenAI's Code Interpreter, can help the model compute and run code. If a task can be done better or faster with a tool other than a language model, use it to get the best result.
Tips:
Use a special search method to find information quickly and properly.
Use a program to do calculations more accurately or to connect to other systems.
Give the model access to specific functions.
6. Test changes systematically
Improving performance is easier if you can measure it. Sometimes a small change in a task may work better for a few examples, but worse for a larger, more representative group of examples. Therefore, to be sure that a change really is an improvement, it is sometimes necessary to do a comprehensive test (also known as an "evaluation").
Tip:
Evaluate the model's answers by comparing them with the best, correct answers.
The above techniques come from the OpenAI "Prompt engineering guide. Check out this guide if you want to dive into this in more detail.
Extra: Create custom instructions in ChatGPT
ChatGPT offers the ability to enter your own instructions and background information about yourself in the "Settings," which contributes to better results.
Click on the '...' by your profile, then choose 'Custom instructions' ('Custom instructions'). Activate this option, and enter information about yourself at the top. Below that, give instructions on how you want ChatGPT to respond. For example, you can specify that you want 'informal text', 'replies in Dutch', and 'detailed replies'.
Take a leap forward in your marketing AI transformation every week
Every Friday, we bring you the latest insights, news and real-world examples on the impact of AI in the marketing world. Whether you want to improve your marketing efficiency, increase customer engagement, sharpen your marketing strategy or digitally transform your business, "Marketing AI Friday" is your weekly guide.
https://nxtli.com/wp-content/uploads/2024/01/6-technieken-voor-effectieve-prompts-NXTLI.png28321594Michiel Schoonhovenhttps://nxtli.com/wp-content/uploads/2019/10/NXTLI_Next_Level_Impact_Logo-1-300x123.pngMichiel Schoonhoven2024-01-04 18:53:082024-05-10 15:59:28Apply these 6 effective prompts techniques for better AI results
Last year we experienced the first phase of awareness and exploration in the world of AI, laying the foundation for the AI era. We have watched a legion of new AI companies launch and existing companies begin to integrate AI into their products and services. In both our private and working lives, we are also discovering AI applications.
By 2024, this development is going to accelerate further. Most companies are going to get serious about integrating AI into their work processes, products and services. We will start to see large-scale experimentation and adoption, with AI becoming prominent in more and more facets of our lives.
Let's see what developments we can expect in 2024 and how they will impact our daily lives.
AI models are getting smarter, more powerful and better
The impact of AI on marketing
The emergence of a team of AI assistants
Access to proprietary language models
New roles and functions in the AI era
Apple introduces AI Diary, a harbinger of things to come
Large Language Models (LLMs) are hallucinatory dream machines
One thing is certain: this time next year, our way of living and working will be radically different. But how exactly? Let's find out together.
AI models are getting smarter, more powerful and better
AI models are evolving rapidly and getting smarter, more powerful and better by the day. We are seeing a shift from traditional language models, which are purely text-based, to more sophisticated multimodal AI models. This new generation of models integpricing images, audio, video and code in addition to text. An example of this is Google's recently announced Gemini.
It is expected that ChatGPT-5, OpenAI's successor to GPT-4, will also be a multimodal model and will be launched early next year. There is speculation that GPT-5 will have more advanced features, including some degree of self-awareness and self-correction, but this is still unconfirmed.
This trend represents not only a shift in how AI affects SEO, but more importantly how it changes user behavior. The zero-click answers of AI chatbots offer an instant solution without the need to navigate through links, ads, pop-ups, SEO-optimized text and cookies. And you won't be retargeted. This is undoubtedly going to have a big effect on traditional search engines and website visits.
The impact of AI on marketing
The impact of AI on our marketing efforts is going to be profound. With advanced AI tools at our fingertips, content creation, distribution and customer service, among others, will not only become easier, but also significantly more powerful.
Let's look at 6 marketing areas where AI can be used.
1. Content creation For many marketers, content creation is the first area they use AI for. They use AI to create (more effectively) content such as text, images, audio and video. The quality of this is going to keep improving, where it is barely, if at all, distinguishable from real.
2. Content distribution AI can help us improve visibility on various channels, such as social media, websites, web shop and email marketing. For example, consider an AI tool that adapts content on your website based on the visitor.
3. Online advertising By using AI for online campaigns, we can analyze huge data sets and learn from user interactions, making our ads more relevant and effective.
4. Customer service AI-powered chatbots offer a new dimension to customer service, allowing us to respond to customer inquiries quickly and efficiently, at any time of day.
5. Content strategy By 2024, AI will be indispensable in developing content strategies. Recognizing trends and analyzing behavior will provide us with deeper insights than ever before.
6. Data and insight AI helps us process mountains of data and gather valuable insights that can strengthen our content marketing strategy and execution.
Wondering what this looks like in practice? Check out a detailed breakdown of the specific marketing aspects where AI can play a rolehere.
The emergence of a team of AI assistants
The two most advanced AI models are OpenAI's ChatGPT-4 and Google's Gemini Ultra. These models are currently superior to all other AI models in the world, which do not get beyond the level of ChatGPT-3.5.
In general, the larger the AI model (which is trained on more data and thus requires more expensive computing power to train. For example, training GPT-4 cost more than $100 million!), the "smarter" it is.
In the world of AI, however, we see a difference emerging between generalist and specialized models. The generalist AIs, such as GPT-4 and Gemini, can handle a wide range of tasks. In contrast, the specialized AIs excel at specific tasks, although they are less flexible.
Imagine a team of AI assistants at your disposal. You have general assistants for many different tasks and more specialized assistants for very specific tasks. These AIs not only work for you, but also with each other. They can delegate tasks to each other, depending on who is best suited for a particular task.
For example, if you want to book a trip, a generalist AI assistant can find the perfect destination based on your preferences. It then engages a specialized AI to handle the booking.
The future of AI is not just about using a single smart assistant, but an entire network of AIs working together. This network will consist of both smart, generalist AIs and less sophisticated but specialized assistants.
Curious about how you can prepare your organization for this future?
Language models transform the way we interact with technology. These models fall into two categories: "closed source," such as ChatGPT-4 and Google's Gemini, where the source code is not publicly available and the owner retains complete control; and "open source," such as Meta's LLaMA. The open source models, share their source code and training data, making them accessible to all.
With the availability of open source language models (Large Language Models -LLMs), the world is open to smart developers to create effective and specialized AIs. These AIs can help with tasks that previously required complex and expensive programs. We are at the dawn of an era in which language models that respond to and learn from their environment will become the norm, both in our work and in our daily lives.
An interesting new player in this field is Mistral AI. This Paris-based startup, founded by former employees of Google's DeepMind and Meta, recently raised an impressive $415 million. With a valuation of $2 billion in just 10 months! (yes, you read correctly), Mistral AI looks set to become a major player. They focus on developing open, decentralized and accountable AI technology.
Mistral AI has also played an important role in the discussion surrounding the European Union's AI legislation and recently launched a developer platform in beta.
New roles and functions in the AI era
As we enter the AI era, many entrepreneurs are asking the question, "How do I ensure the right capacity and people to make our transition to AI happen?"
As your AI use m atures, the structure of your team must evolve. This means new roles and functions are needed:
Core function:
Chief AI Officer: Responsible for overarching AI strategy, policy, and implementation. Important is a thorough knowledge of business strategy, change management, and marketing technology procurement.
Must-have roles:
AI Project/Program Manager(s): Drives specific AI projects and programs. Functions as the central hub of the organization with additional AI representatives in each department.
Optional rollers:
Integration expert: Specializes in integrating AI tools into existing processes and systems. AI developers: For building and maintaining custom AI tools.
Key stakeholders:
Marketing, Sales, Chief Editors, Analysts, Legal, Corporate Culture, and Data Security are critical to the smooth integration of AI into the organization.
It is essential to assemble a versatile team that can harness and apply the capabilities of AI in all aspects of the organization. With an AI steering committee, you can structure and organize this well. The future belongs to companies that embrace these new roles and fully harness the power of AI.
Apple introduces AI-powered Journal
Apple says it is working on advanced AI language models for iPhones and other devices, and has made significant breakthroughs in their implementation and storage. The company's researchers have developed a technique to store AI data in flash memory, making it possible to run complex AI assistants and chatbots on devices with limited memory.
These developments could lead to improved Siri features, real-time translation, and advanced capabilities in photography and augmented reality. Apple aims to integrate AI into as many apps as possible, while focusing on privacy and responsible AI development.
The launch of the AI-powered Diary app, may be a harbinger of innovative AI applications on the Iphone and other devices.
The Diary app can automatically suggest key moments from your recent locations visited, photos, music and workouts, to help you get started writing. With the update to OS 17.2, the Diary app will be added automatically, fully integrated into the Apple ecosystem.
Apple emphasizes that privacy is central to the development of this app. Data in the Diary app is encrypted while your iPhone is locked, and users have the option to activate additional security measures such as passcode, Face ID or Touch ID. All diary data is end-to-end encrypted and stored in iCloud, so only you have access to your memories.
Interestingly, diary suggestions are generated locally on the device. This means you have complete control over which suggested moments are shared with the app and added to your journal.
LLMs are hallucinatory dream machines
In closing, I share a fascinating insight from Andrej Karpathy, a leading AI researcher who also recently gave an impressive presentation on the development of LLMs. He highlights the "hallucination" property of AI such as ChatGPT, which often shows up in answers that seem out of context or incorrect. Many people see this as a flaw, but Karpathy offers a different perspective: LLMs are essentially dream machines.
Karpathy explains that what we think of as "hallucination" is in fact the core function of LLMs: they generate ideas and answers based on their training.
'We direct their dreams with prompts. The prompts start the dream, and based on the LLM's hazy recollection of its training documents, most of the time the result goes someplace useful. It's only when the dreams go into deemed factually incorrect territory that we label it a "hallucination." It looks like a bug, but it's just the LLM doing what it always does.'
Karpathy also makes clear the difference with search engines, which only reuse existing information and cannot create new ideas. An LLM is dreaming 100% and has the hallucination problem. A search engine is dreaming 0% and has the creativity problem. He emphasizes that although hallucination is inherent in LLMs, we need to find ways to control this property in AI assistants.
This vision from Karpathy offers an interesting new perspective on AI, emphasizing the importance of leveraging their unique creative capabilities. Looking ahead to 2024, it is essential to focus on the strengths of AI tools and deploy them where they are most effective for us. By continuously experimenting and regularly testing for improvements, we can make the most of the lightning-fast developing AI technology.
Take a leap forward in your marketing AI transformation every week
Every Friday, we bring you the latest insights, news and real-world examples on the impact of AI in the marketing world. Whether you want to improve your marketing efficiency, increase customer engagement, sharpen your marketing strategy or digitally transform your business, "Marketing AI Friday" is your weekly guide.
https://nxtli.com/wp-content/uploads/2023/12/Een-multi-modal-language-model-in-AI.jpg10241024Michiel Schoonhovenhttps://nxtli.com/wp-content/uploads/2019/10/NXTLI_Next_Level_Impact_Logo-1-300x123.pngMichiel Schoonhoven2023-12-22 20:19:312024-05-10 15:59:28In 2024, AI developments are going to accelerate
Apple says it is working on advanced AI language models for iPhones and other devices, and has made significant breakthroughs in their implementation and storage. The company's researchers have developed a technique to store AI data in flash memory, making it possible to run complex AI assistants and chatbots on devices with limited memory.
Apple Diary
These developments could lead to improved Siri features, real-time translation, and advanced capabilities in photography and augmented reality. Apple aims to integrate AI into as many apps as possible, while focusing on privacy and responsible AI development.
The launch of the AI-powered Diary app, may be a harbinger of innovative AI applications on the Iphone and other devices.
The Diary app can automatically suggest key moments from your recent locations visited, photos, music and workouts, to help you get started writing. With the update to OS 17.2, the Diary app will be added automatically, fully integrated into the Apple ecosystem.
Apple emphasizes that privacy is central to the development of this app. Data in the Diary app is encrypted while your iPhone is locked, and users have the option to activate additional security measures such as passcode, Face ID or Touch ID. All diary data is end-to-end encrypted and stored in iCloud, so only you have access to your memories.
Interestingly, diary suggestions are generated locally on the device. This means you have complete control over which suggested moments are shared with the app and added to your journal.
Take a leap forward in your marketing AI transformation every week
Every Friday, we bring you the latest insights, news and real-world examples on the impact of AI in the marketing world. Whether you want to improve your marketing efficiency, increase customer engagement, sharpen your marketing strategy or digitally transform your business, "Marketing AI Friday" is your weekly guide.
Developments in the field of AI are having a significant impact on traditional search engines such as Google Search and, consequently, website traffic. Search behavior online is changing rapidly with the emergence of AI tools such as ChatGPT and Perplexity, founded by OpenAI researcher Aravind Srinivas. The "even Googling" is giving way to "even GPTing. What does that mean for you?
With tremendous dominance in search, 9 out of 10, and ad revenue of more than $280 billion last year, Google faces a significant challenge.
My personal experience since the mid-2023s, using AI exclusively as a starting point for my searches (from general information to very specific system user instructions), has caused me to virtually stop using Google Search. I find the user experience with AI chatbots much nicer and more effective. I hear and read from more and more people that they feel the same way.
This trend represents not only a shift in how AI affects SEO, but more importantly how it changes user behavior. The zero-click answers of AI chatbots offer an instant solution without the need to navigate through links, ads, pop-ups, SEO-optimized text and cookies. And you won't be retargeted. This is undoubtedly going to have a big effect on traditional search engines and website visits.
Understanding and anticipating AI-driven search behavior
It is important to understand the shift to AI-driven search behavior and respond proactively to it. Therefore, consider the following:
Current search behavior of your target audience: Analyze how your target audience currently searches, what tools they use to do so and what their preferences are. Repeat this regularly.
Your website's future role: Define the desired role your website plays in a landscape where AI provides immediate answers. How will your site remain relevant and attractive to your visitors?
Adapting your 2024 strategy: Consider whether your strategy adequately addresses the changes in search behavior and the way content is consumed by your audience. This may require a thorough rethinking of your current approach.
Creating AI-friendly and effective content
The goal is to develop content that not only scores well in the new "Search Generative Experience (SGE)" - generating direct, integrated answers to search queries using AI, but also effectively drives visitors to your website.
This includes:
Visibility in AI tools: Make sure your content is visible and attractive to AI tools such as ChatGPT. This means the content should be very distinctive, clear and informative for answering questions asked.
Diversification of content formats: Video and audio content, such as podcasts, will become more important. These formats offer a dynamic and more distinctive way to reach and engage your audience.
Focus on own experiences and unique insights: Share your own experiences, conduct relevant research, and offer unique insights that are also interesting for other media and platforms to publish. This type of content is attractive to AI because it offers unique perspectives and in-depth analysis.
This allows you to respond effectively to the changing dynamics of search engines and AI-driven search behavior. You also simultaneously ensure the relevance and visibility of your content.
Preparing for the future
What actions can you take to prepare for the zero-click world:
Establish an AI steering committee: Create a multidisciplinary team to lead the development and implementation of AI strategies in your organization. This group will act as "discoverers" of new AI capabilities.
Experiment with chat search functionality from Bing and Bard: This provides initial insights into the development of generative search experiences. You can also see the extent to which you yourself are or are not visible on key search terms.
Analyze risks of declining website traffic: Estimate the impact if traffic from Google Search (and paid search) declines starting in 2024.
Determine your ultimate area of expertise: Research which search terms you have a strong position on that you want to protect (the questions you are the very best at answering) and which questions you want to build a dominant position with answers on.
Experiment with other channels: Reduce reliance on SEO and SEA by experimenting with other channels where your target audience is active and expects and accepts you.
The importance of content
While content remains crucial to being found, it is important not to focus on producing more generic content with AI. Instead, a thoughtful and transformative content strategy is required.
We are ready to support you in this and together develop a strategy that fits the new reality of search behavior and content consumption.
Take a leap forward in your marketing AI transformation every week
Every Friday, we bring you the latest insights, news and real-world examples on the impact of AI in the marketing world. Whether you want to improve your marketing efficiency, increase customer engagement, sharpen your marketing strategy or digitally transform your business, "Marketing AI Friday" is your weekly guide.
https://nxtli.com/wp-content/uploads/2023/12/Welkom-in-the-nul-klik-wereld.jpg10241022Michiel Schoonhovenhttps://nxtli.com/wp-content/uploads/2019/10/NXTLI_Next_Level_Impact_Logo-1-300x123.pngMichiel Schoonhoven2023-12-15 14:47:442024-03-15 19:30:52AI and the impact on search engines: Welcome to the zero-click world
