Velocity Ascent

Looking toward tomorrow today

AI

AI Biomechanical Engines: powering real-world solutions.

· ·

“AI-generated biomechanical engines” refers to a conceptual or technological framework that integrates artificial intelligence (AI) with biomechanical systems to create dynamic, adaptable mechanisms or devices.

That sure is a mouthful 🙂

Putting it simply, these engines could be a blend of biological principles (like human or animal movement) and mechanical or robotic engineering, powered or optimized through AI algorithms. Here’s a breakdown of the components:

1. AI (Artificial Intelligence): AI in this context would involve algorithms and computational models that enable machines or systems to learn, adapt, and make decisions, often based on data inputs. This could include machine learning, neural networks, or deep learning techniques, which allow systems to improve over time or adapt to new conditions.

2. Biomechanical: This term refers to the study and application of mechanical principles to biological systems. It typically involves understanding how living organisms, such as humans or animals, move, function, and interact with their environments. In the context of “engines,” it could refer to mechanical systems that replicate or augment biological movement, such as in prosthetics, robotics, or exoskeletons.

3. Engines: In this context, “engines” likely refers to systems or machines that drive or power a mechanism. It can include anything from the propulsion systems in robots to more complex devices designed to mimic biological functions, such as limbs, joints, or circulatory systems.

Practical Examples:

• Robotics: Robots that use AI to adapt their movements based on real-time data from their environment, effectively mimicking human or animal biomechanics. For instance, robots that use AI to improve their walking gait based on biomechanical principles.

• Prosthetics & Exoskeletons: AI-driven prosthetics or exoskeletons that use machine learning to optimize the movement of artificial limbs in ways that mimic natural human motion. These devices could learn to respond to a user’s intent more fluidly, adjusting for different terrains or activities.

• Biohybrid Systems: Devices that combine both biological and mechanical components, powered by AI to enhance movement or function. An example might be biohybrid robots that incorporate living cells or tissues, with AI systems guiding their movement.

Future Implications:

In the future, AI-generated biomechanical engines could lead to more advanced prosthetics, adaptive robotic limbs, and even new forms of biological augmentation.

Such systems could be used in medicine, personal assistance, military applications, or even in enhancing human capabilities. By integrating AI and biomechanics, these systems might be able to respond to their environment in real-time, providing more natural, human-like, or even superhuman abilities.

Examples in current practice

Certainly! Here are the links to the relevant sources for the examples I mentioned:

1. OpenAI’s Dactyl

• Overview of Dactyl:

• OpenAI’s Robotics: Solving Rubik’s Cube with a Robotic Hand (2018)

• OpenAI’s Dactyl

Dactyl uses reinforcement learning to manipulate objects like the Rubik’s Cube, demonstrating the combination of AI and biomechanical robotics.

2. Boston Dynamics’ Atlas

• Boston Dynamics Atlas:

• Atlas Robot – Boston Dynamics

• Atlas’ Parkour Performance (Video demonstration)

Atlas is known for its ability to perform complex physical tasks like running, jumping, and backflips using advanced machine learning algorithms to simulate human-like biomechanics.

3. Honda’s ASIMO

• Honda ASIMO:

• ASIMO – Honda Worldwide

• ASIMO’s Capabilities and Developments (ASIMO demo video)

ASIMO is a humanoid robot that combines AI with biomechanics to perform human-like movements, from walking to dancing, and adapt to its environment in real time.

4. ExoAtlet Exoskeleton

• ExoAtlet – Exoskeleton:

• ExoAtlet Official Website

• ExoAtlet Rehabilitation Exoskeleton (YouTube video)

The ExoAtlet exoskeleton is designed to help people with mobility impairments regain movement and stability using AI to adapt to the user’s biomechanics.

5. Sharp’s RoBoHoN

• RoBoHoN by Sharp:

• RoBoHoN Official Page – Sharp

• RoBoHoN in Action (YouTube)

RoBoHoN is a humanoid robot that integrates AI to enable complex movements like walking and conversation. It is a smaller, more interactive robot with applications in communication and mobility.

These resources provide additional context and examples of how AI has been applied to biomechanical engines, including robots and exoskeletons.

Generative AI: Revolutionizing Video & Image Production

· ·

Video & Imagery Production is undergoing a seismic shift, thanks to the arrival of advanced AI technologies like Midjourney and Runway.

The workflow of video and image production is undergoing a seismic shift, thanks to the merging of advanced AI technologies like Midjourney v6.1 and Runway Gen-3. This powerful combination is more than just a technical marvel; it’s a transformative tool for marketing agencies, revolutionizing the way they approach video content creation.

Accelerating Production with AI Synergy

For marketing agencies, the need to produce high-quality video content quickly and efficiently is paramount. The integration of Midjourney v6.1 and Runway Gen-3 offers a solution by streamlining the pre-production process. These tools allow creators to visualize and animate storyboards at an unprecedented pace, reducing the time it takes to move from concept to final product.

This speed and flexibility mean that agencies can handle more projects simultaneously, increasing their output without sacrificing quality. By automating time-consuming tasks, AI enables teams to focus on refining creative ideas rather than getting bogged down in the technical aspects of production.

Our Experimentation:

We conducted several dozen tests and reviewed numerous articles and papers online to evaluate the integration of Generative AI in early-stage storyboard development for videos and its impact on agency workflows. The results were eye-opening.

Source Image Creation:

For this experiment, the creative brief required the following: the image must include a group of friendly seniors smiling around a smartphone at a picnic table in the summer, during golden hour. The first attempts were promising, and after a few subtle adjustments to the prompt, we began to achieve ‘realistic’ results close to stock photography.

Early experiment generating a group of Seniors.

A good start, but the image looked a bit too ‘Princess Bride,’ with colors that were far too saturated. While it’s possible to use the SREF in Midjourney and a color palette from Adobe Color to color grade your Midjourney imagery, we opted for a more hands-on approach. We created a custom color range tailored to our target, resulting in improved color balance and a higher degree of ‘believability.’ Much of the effort went into refining the output.

Array of images generated during refinement.

Our second goal was to enhance the appearance of the seniors. Initially, they appeared, forgive the ageism, too old for our demographic requirements. Finally, we adjusted the prompt to eliminate unwanted artifacts, such as seniors holding two phones instead of one and glasses melting into the table.

Final image chosen for use in video storyboard.

The final image did more than achieve the creative brief baseline requirements – it created a “feeling”. You can almost feel the warmth of a late summer day and the joy of the gathered Seniors. With the image and prompts finalized our next step was to move over to Video Gen.

Video Generation:

Since we would just be using AI Video Generation as a storyboard and not for broadcast we ran the first series as “vanilla prompts”. This allowed us to quickly generate action that could be placed in the storyboard sample with lots of room for feedback and adjustment.

Early stage test of AI Generative Video

Our initial tests yielded the expected otherworldly creations; the video starts off normally but quickly spirals into a surreal, science fiction fever-dream with all but one of the Seniors sliding into oblivion. Applying a revised prompt set and rethinking the scene resulted in a new direction more in line with the requirements.

Late stage test of AI Generative Video

The results are nothing short of fascinating, the motion is smooth and the Seniors appear to be in a natural setting acting normally. There are still some anomalies that are fine-tuned out of the finished video. For example in the sample video above you will notice on the left side of the scene the smartphone morphs into the pint glass. A few more edits to the prompt will resolve this.

Given the time and resources typically needed for custom video production or extensive research across stock agencies, the speed of this process is remarkably faster. Even more exciting is that we are only at the dawn of this technology.

Cost Reduction and Enhanced Creative Output

The financial benefits of this AI-driven approach are significant. By reducing the time and resources required for video production, agencies can lower costs for their clients while delivering even more compelling content. This efficiency opens up opportunities for smaller businesses to access high-quality video production, which was previously only within reach of larger companies with bigger budgets.

Moreover, the enhanced creative capabilities of these AI tools allow for rapid prototyping and experimentation. Agencies can quickly test and iterate on ideas, ensuring that the final product aligns perfectly with the client’s vision. This ability to fine-tune creative concepts on the fly is a game-changer for marketing campaigns, where the margin for error is often slim.

While AI can automate certain aspects of video creation, it lacks the nuanced understanding of brand voice, audience preferences, and emotional resonance that only human creators can bring.

The Human Element: Essential and Irreplaceable

Despite the incredible advancements in AI technology, the human element remains crucial in the production process. While AI can automate certain aspects of video creation, it lacks the nuanced understanding of brand voice, audience preferences, and emotional resonance that only human creators can bring.

The role of the human in this AI-driven workflow is to guide and shape the creative direction, ensuring that the content not only meets technical standards but also connects with viewers on a deeper level. AI tools serve as powerful assistants, augmenting human creativity but never replacing it. This collaboration between human ingenuity and machine efficiency is where the true potential of AI in video production lies.

Transforming the Advertising and Entertainment Industries

The impact of this AI merging extends beyond marketing agencies to the broader advertising and entertainment industries. The ability to rapidly prototype ideas and create high-quality visual content democratizes the production process, making it accessible to a wider range of creators, from seasoned professionals to enthusiastic hobbyists.

As this technology continues to evolve, we’re likely to see a shift in how visual media is consumed and interacted with. The lines between imagination and realization are becoming increasingly blurred, allowing for more immersive and personalized storytelling experiences.

Impact on Traditional Video Production:

AI tools are transforming video production by automating many aspects of the process, such as editing, special effects, and even scriptwriting. This automation can greatly speed up production timelines and reduce costs, enabling agencies to produce more content with fewer resources. For smaller teams or independent creators, AI provides access to high-quality production tools that were once out of reach.

However, this increased efficiency could lead to a reduction in the demand for certain roles, such as video editors, animators, and even directors. As AI becomes more sophisticated, the need for manual intervention in repetitive or technical tasks will decrease, potentially leading to job losses in these areas.

Impact on Stock Agencies:

AI-generated content also poses a challenge to traditional stock agencies. With AI tools capable of creating high-quality images, videos, and even audio, the reliance on stock libraries may diminish. Creators can now generate custom content tailored to their specific needs without having to sift through existing libraries. This shift could reduce demand for traditional stock footage and images, impacting the revenue streams of stock agencies.

Potential Job Losses:

While AI has the potential to displace certain jobs, it’s important to recognize that it also creates new opportunities. Jobs focused on AI tool development, data analysis, and AI integration into creative processes will likely grow. Additionally, roles that require a high degree of creativity, strategic thinking, and emotional intelligence—skills that AI currently cannot replicate—will remain essential.

The key for professionals in the industry will be to adapt to these changes by learning to work alongside AI, using it as a tool to enhance their capabilities rather than viewing it as a replacement. Those who can harness the power of AI to augment their creativity and productivity will find new opportunities in the evolving landscape of video production and stock media.

The importance of the human touch in this process cannot be overstated.

The Future of Storytelling

The future of video production is here, and it’s powered by AI. For marketing agencies, embracing this technology means not only staying competitive but also pushing the boundaries of what’s possible in creative content creation.

However, the importance of the human touch in this process cannot be overstated. As we move forward into this new era of AI-driven production, the collaboration between human creativity and AI efficiency will be the key to unlocking new levels of innovation and storytelling.

The Artistry of Prompt Engineering

· ·

At its core, prompt engineering embodies a blend of technical precision and creative writing finesse.

In the realm of artificial intelligence (AI), prompt engineering stands as a pivotal technique, wielding immense power in shaping the capabilities and behaviors of AI models. This post examines the artistry, potential and potency of prompt engineering, focusing on a variety of prominent platforms: ChatGPT, Davinci, Haiper, and Midjourney.

A prompt is natural language text describing the task that an AI should perform.

As we unravel the intricacies of this programming language, we’ll uncover how it empowers users to mold AI outputs to suit diverse needs and purposes.

Understanding Prompt Engineering:

Prompt engineering is the process of structuring an instruction that can be interpreted and understood by a generative AI model.

Prompt engineering revolves around crafting precise instructions or prompts that guide AI models in generating desired outputs. These prompts serve as the input for AI systems, influencing the content, tone, and style of their responses. Through careful crafting, users can steer AI towards generating outputs that align with specific objectives or criteria.

Chain-of-thought (CoT) prompting

Introduced in Wei et al. (2022), chain-of-thought (CoT) prompting enables complex reasoning capabilities through intermediate reasoning steps. The technique is aimed at enhancing the reasoning ability of large language models (LLMs) by guiding them through a problem-solving process in a series of intermediate steps before arriving at a final answer. This approach mimics a train of thought, facilitating logical reasoning and addressing challenges in tasks requiring multi-step solutions, such as arithmetic or commonsense reasoning questions. For instance, when presented with a question like “The cafeteria had 23 apples. If they used 20 to make lunch and bought 6 more, how many apples do they have?” a CoT prompt could guide the LLM to break down the problem into sequential steps, leading to a comprehensive answer.

Chain-of-thought prompting enables large language models to tackle complex arithmetic, commonsense, and symbolic reasoning tasks- reasoning processes are highlighted. Image Source: Wei et al. (2022)

Initially, CoT prompts included a few question-and-answer examples, making it a few-shot prompting technique. However, the effectiveness of CoT has been demonstrated with the addition of simple prompts like “Let’s think step-by-step,” transitioning it into a zero-shot prompting technique and enabling easier scalability without the need for numerous specific examples. When applied to PaLM, a 540B parameter language model, CoT prompting significantly improved its performance on various tasks, achieving state-of-the-art results on benchmarks such as the GSM8K mathematical reasoning benchmark. Further enhancements can be achieved by fine-tuning models on CoT reasoning datasets, which could lead to improved interpretability and performance.

Platform Agnostic: Exploring Prompt Engineering on ChatGPT, Davinci, Haiper, and Midjourney

There currently exists a myriad of tools and frameworks designed to harness the power of artificial intelligence. Among these, ChatGPT and Midjourney stand out as prominent examples, each offering unique capabilities and applications.

ChatGPT: AI Copywriting

ChatGPT, developed by OpenAI, stands as a pioneering platform in the domain of conversational AI. Prompt engineering plays a fundamental role in shaping the interactions facilitated by ChatGPT. By crafting tailored prompts, users can steer conversations in desired directions, maintain coherence, and achieve specific conversational goals.

Techniques in Prompt Engineering for ChatGPT:

  1. Contextual Prompting: Leveraging contextual cues within prompts to provide relevant information and guide the AI’s understanding of the conversation’s flow.
  2. Persona Establishment: Crafting prompts that establish a consistent persona for the AI, shaping its tone, demeanor, and overall personality.
  3. Prompt Refinement: Iteratively refining prompts based on AI responses to achieve optimal conversational outcomes.
121 separate runs of /imagine a sailboat --sref random

Midjourney: AI Graphics

AI Art Generator From Text

Midjourney emerged early as a novel platform in the AI landscape, offering tools and frameworks for prompt engineering tailored towards diverse applications. With a focus on narrative generation and storytelling, Midjourney empowers users to craft compelling narratives through strategic prompt design.

Techniques in Prompt Engineering for Midjourney.

Narrative Structure, Emotional Contextualization and Dynamic Prompting

Narrative

Narrative Structure: Designing prompts that outline the desired narrative arc, including key plot points, character interactions, and thematic elements.

Emotion

Emotional Contextualization: Incorporating emotional cues and context within prompts to evoke specific feelings or reactions from the AI-generated narrative.

Dynamic

Dynamic Prompting: Employing dynamic prompts that adapt based on AI-generated content, ensuring coherence and narrative continuity.

slide 1
Image Slide 2
Image Slide 1
Image Slide 3
Image Slide 3
Image Slide 3
Image Slide 3
Image Slide 3
Image Slide 3
Image Slide 3
Image Slide 3
previous arrowprevious arrow
next arrownext arrow

In the example above we see the progression of the art though adjustment and refinement. Each slide represents the process of engineering the prompts, for example “PROMPT: Midjourney – young boy, age 5-6 years old, wearing blue helmet, holding steering wheel, in red 1952 Murray Champion Pedal Car rolling toward camera, smiling, natural lighting, Nikon D850 28mm, global illumination –ar 16:9 –v 6.0” Prompts can be very descriptive whether it is choosing an actual real world camera or a time of day. The end result can be very compelling – will it replace real world photography for example. I think in some cases where quick imagery is needed the tool creates passable art. Venturing into highly detailed brand-accurate art is still evolving.

garbage in, garbage out (GIGO)

On two occasions I have been asked, “Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?” … I am not able rightly to apprehend the kind of confusion of ideas that could provoke such a question.

— Charles Babbage, Passages from the Life of a Philosopher[5]

Haiper: AI Video

Haiper emerges as a groundbreaking advancement in the realm of artificial intelligence, offering a potential trajectory towards achieving Artificial General Intelligence (AGI). Its distinctive feature lies in its utilization of a unique perceptual foundation model—a feat achieved by only a select few in this domain.

In a fresh take on AI initiatives, Haiper’s approach is rooted in a philosophy that prioritizes not only technological prowess but also community collaboration and creative synergy. Founded by distinguished alumni from industry giants like Google DeepMind and TikTok, as well as leading research labs in academia, Haiper is working to blend next-gen machine learning with a refined perspective on creativity.

This innovative approach potentially positions Haiper as more than just another AI tool. It is a versatile creativity platform that breaks with traditional industry boundaries, placing emphasis on fun, shareability, and community engagement.

Davinci: AI Graphics & Video (Multisource)

AI Art Generator From Text

DaVinci features the latest state-of-the-art AI technology to generate unique artworks and photorealistic images. It offers various AI models to choose from, including its own custom AI model, DaVinci XL, Stable Diffusion, DALL·E 3, and Midjourney.

The newest release, DaVinci Resolve 19, adds two new AI features that make video editing more efficient: the IntelliTrack AI point tracker for object tracking, stabilization and audio panning, and UltraNR, which uses AI for spatial noise reduction.

The Artistry of Prompt Engineering:

At its core, prompt engineering embodies a blend of technical precision and creative writing finesse. It requires an understanding of AI capabilities, linguistic nuances, and user objectives. Crafting effective prompts entails a deep appreciation for language, narrative structure, and contextual subtleties, elevating it to an art form in its own right.

Prompt engineering serves as a cornerstone in the development and deployment of AI systems, enabling users to wield unprecedented control over AI-generated outputs. As AI continues to evolve, the mastery of prompt engineering will remain indispensable, unlocking new frontiers in human-AI collaboration and creativity.

Who was the first prompt engineer?

Pinpointing the exact “first” prompt engineer in the context of AI is a bit challenging, as prompt engineering has evolved over time with the development of AI technologies. However, we can attribute the early origins of prompt engineering to researchers and developers who explored techniques to influence the behavior of AI systems through tailored inputs.

In the realm of natural language processing and early chatbots, developers experimented with crafting prompts or inputs to elicit specific responses from AI models. For example, in the 1960s and 1970s, Joseph Weizenbaum created ELIZA, one of the earliest chatbots, which relied on pattern-matching techniques to simulate conversation. While not exactly prompt engineering in the modern sense, Weizenbaum’s work laid the groundwork for manipulating interactions with AI systems through carefully designed inputs.

The most famous script, DOCTOR, simulated a psychotherapist of the Rogerian school (in which the therapist often reflects back the patient’s words to the patient),[9][10][11] and used rules, dictated in the script, to respond with non-directional questions to user inputs. 

As AI technologies advanced, particularly with the rise of deep learning and large language models like GPT (Generative Pre-trained Transformer), prompt engineering gained prominence as a method for fine-tuning and controlling AI-generated outputs. Researchers, developers, and practitioners across academia and industry contributed to the development and refinement of prompt engineering techniques, shaping its evolution into a sophisticated discipline.

So, while there may not be a single “first” prompt engineer, the concept emerged gradually as AI technologies evolved, with contributions from various individuals and communities.

An uncritical embrace of technology?

The “uncritical embrace of technology” refers to a phenomenon where individuals, or society at large, enthusiastically adopt and rely on technological advancements without adequately considering their potential drawbacks, limitations, or broader societal implications. This uncritical acceptance often stems from the perceived benefits or conveniences offered by technology, leading to a lack of critical reflection on its long-term effects.

In conversational interfaces such as ChatGPT or narrative generation platforms like Midjourney users may enthusiastically embrace these technologies for their convenience, entertainment value, or utility in various applications. They may appreciate the ease of generating conversational content or narratives using AI-powered platforms, without necessarily critically examining the underlying algorithms or potential biases in the generated outputs.

Similarly, developers and organizations may prioritize the development and deployment of conversational AI and narrative generation tools to meet market demand, improve user experiences, or achieve specific business objectives. In doing so, they may focus more on technical innovation and functionality rather than thoroughly evaluating the ethical implications or societal impacts of these technologies.

Several factors contribute to the uncritical embrace of technology:

  1. Techno-optimism: Many people hold a belief in the inherent goodness or progressiveness of technology, viewing it as a solution to various problems and a driver of societal advancement. This optimism can lead to a bias towards embracing new technologies without fully evaluating their potential risks.
  2. Market-driven innovation: In a competitive market environment, there is often pressure for companies to continuously innovate and release new products or services. This drive for innovation can prioritize speed and novelty over thorough consideration of ethical, social, or environmental implications.
  3. Convenience and efficiency: Technology often promises to streamline tasks, improve efficiency, and enhance convenience in various aspects of life. As a result, individuals may readily adopt new technologies without questioning their broader impacts, focusing instead on immediate benefits.
  4. Social influence and peer pressure: Social norms and peer influence can play a significant role in shaping attitudes towards technology. If a particular technology becomes widely adopted or socially endorsed, individuals may feel compelled to embrace it without questioning its implications.
  5. Limited understanding: Not everyone possesses a deep understanding of the underlying mechanisms or implications of technology. As a result, individuals may accept technological innovations at face value, without fully grasping their potential consequences.

The uncritical embrace of technology can have several consequences, including:

  • Ethical dilemmas: Technologies may raise ethical questions related to privacy, surveillance, autonomy, and fairness, which may not be adequately addressed if adoption is uncritical.
  • Social impacts: The rapid adoption of technology can lead to societal changes that may exacerbate inequalities, disrupt traditional industries, or alter social norms and behaviors.
  • Environmental concerns: Some technologies may have negative environmental impacts, such as increased energy consumption, resource depletion, or pollution, which may be overlooked in the pursuit of innovation.

To mitigate the risks associated with the uncritical embrace of technology, it’s essential to promote critical thinking, ethical considerations, and inclusive decision-making processes in the development, deployment, and regulation of technology. This approach can help ensure that technological advancements are aligned with broader societal values, goals, and well-being.

Sources:

Here are sources and links where you can find more information on the subjects of technology ethics, societal implications of technology, and critical thinking:

Technology Ethics:

  1. The Markkula Center for Applied Ethics – Technology Ethics: This center, affiliated with Santa Clara University, offers a wealth of resources on technology ethics, including articles, case studies, and research papers.
    Website: https://www.scu.edu/ethics/ethics-resources/ethical-decision-making/technology-ethics/
  2. IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems: The IEEE Global Initiative provides guidelines, reports, and resources on the ethical development and deployment of autonomous and intelligent systems.
    Website: https://ethicsinaction.ieee.org/

Societal Implications of Technology:

  1. Pew Research Center – Internet & Technology: Pew Research Center conducts surveys and studies on the impact of technology on society, covering topics such as digital privacy, online behavior, and the future of work.
    Website: https://www.pewresearch.org/internet/
  2. MIT Technology Review: MIT Technology Review provides in-depth analysis and reporting on emerging technologies and their societal impacts, including articles on AI ethics, data privacy, and digital transformation.
    Website: https://www.technologyreview.com/

Critical Thinking:

  1. The Foundation for Critical Thinking: This organization offers resources and materials to promote critical thinking skills, including books, articles, and online courses.
    Website: https://www.criticalthinking.org/
  2. The Critical Thinking Community: The Critical Thinking Community provides educational resources and tools for fostering critical thinking skills in both academic and professional settings.
    Website: https://www.criticalthinking.org/pages/about-the-critical-thinking-community/858

Additional Resources:

  1. Stanford Encyclopedia of Philosophy – Philosophy of Technology: The Stanford Encyclopedia of Philosophy offers an overview of the philosophy of technology, covering topics such as technological determinism, ethics, and social impacts.
    Website: https://plato.stanford.edu/entries/technology/
  2. Center for Humane Technology: The Center for Humane Technology advocates for the ethical design and use of technology to promote well-being and human flourishing. Their website features articles, podcasts, and resources on topics related to digital well-being and technology addiction.
    Website: https://www.humanetech.com/

These sources provide valuable insights and perspectives on the ethical, social, and cognitive dimensions of technology, empowering individuals to engage critically with the challenges and opportunities presented by technological advancements.