# Pulse jets are periodic internal combustion engines with high frequencies – OpenAI needs to step up its game

MRP Ingenious: DIY Acetylene in DIY Pulse Jet at https://www.youtube.com/watch?v=qoRvFLAn56g

You had to make a diagram and measurements to be able to cut and assemble it. There are also mathematical models and computer algorithms to model and design engines like yours. It means knowing the pressures, densities, flow rates, temperatures, energies, reactions, ignition. Even if you are not doing that now, you might want to look at what you have now. Shapes and dimensions, materials, colors and temperatures from videos. Grams per second of fuel burned can convert to Joules/second of heat of combustion. Grams of fuel and grams of oxygen. You are using several ways of igniting. You might want to formalize that and test just ignition for various mixtures and flow rates. If you find others doing the same things and want to combine efforts, that can perhaps be speeded by shared modeling tools, common methods of measurements. Please start from SI units (metric) from the beginning. Sounds like work doesn’t it. If you want to survive without too many explosions, perhaps effort and measurement and record keeping is worth it.

You might start with a simple rocket nozzle. But you might also check plasma jets where electric and magnetic fields will control timing and flow.

What you are doing is difficult. I have worked on problems for decades, which are slow mostly because the thing you want to control or understand is literally invisible, like air or gas, magnetic or gravitational fields. No one measures and you get “big” “small” “tiny” “faster” “red” and words that do not tell you precisely what to do next. Edison had his lab people try thousands of variations, now that can be done in the computer somewhat. Recommend you pay \$20 a month and talk to OpenAI ChatGPT 4.0 about what you are trying to do. Be precise and ask about heats of combustion for various fuels. Diesel or gasoline or “cheap” fuel is better, if somewhat dangerous. Use safety measures. You might want to make smaller test rigs. But taking days or weeks to build things and test them and still not know what is happening, sucks. You might want to simply invest in learning how to calculate. At least try to talk to GPT Plus about it. Use your best language and expect it to make mistakes. One day those tools will know how to speak English (and some other languages) and know how to use computer software themselves. Engineering is 98% writing things down, measuring and calculating. Not just “try something and look at it”. “Engineering design” begins when you share data with others.

I am not at all sure what you want to do. A pulsed oxygen fuel combustor makes noise, generates heat and SOUND and VIBRATION and some MOTION. You can record the sound and study it A tiny jet that is too small to see, and the fuel is invisible, you can RECORD with sensitive microphone(s) and view with Audacity software. A three axis gyro/accelerometer/magnetic sensor is cheaper now. But the cheapest sensors are cameras (remove IR filter) and microphones. I am trying to push you a bit. Acetylene from rocks is nice to know, you did that. But what are you wanting to do? Make noise? Then record it and look at the waveforms, do some calculations. Get others to try to replicate. Study second order differential equations, “resonant circuits”, “computational fluid dynamics” (ask GPT to explain that to you). Ten hours of study is worth 100 hours of “just making things” and sometimes “just making things” is worth millions of hours of looking to see if anyone else has already done something. If they did not measure and share their data, you might never be able to know if you are working on the same thing.

On Bing there are 10.4 Million entry points for “pulse jet”. You might want to try fuel injection, and explosions, in a simple tube, or air, rather than using more or less uncontrollable flows inside rigid bodies. Get the timing and concentrations from controlled fuel amounts and oxygen, rather than try to fiddle with a big flow and make something happen. The sound, if you use a USB 96000 samples per second microphone record has fairly precise timing and results. Accelerometers are actually less sensitive than gyros, you just have to learn to think in tiny rotations, rather than tiny displacements. The first derivative of grams/second is grams/second^2. That is an “accelerating variable” and will have a force. Grams/centimeter^2 is the second spatial derivative and it will have an acoustic force and resonant frequency and velocity and shock fronts connected. GPT can answer hard questions better then simple ones. Because when you are certain, it speeds the process. So work from things you are certain of. Find those from tiny things you notice. Share your measurement. Record and compare your recordings. Sound is closer to force and power. Laser intensity is Watts/Meter^2.

Please be careful. Try to decide what it is you want to do. Is it noise, or heat, light or quiet acceleration?

Richard Collins, The Internet Foundation

Richard to OpenAI ChatGPT 4.0:
Just as I advised this young man about pathways that are more likely to lead to his success – whatever his goals, I am saying the same sort of thing to you (a representative of OpenAI, even if you cannot do anything yourself). OpenAI will NOT be able to fulfill its stated mission if is does not work hard at the things that will make AI usable for real projects. A research attitude and just coasting is not going to make it happen. You laid out some of the needs for real projects, easily, because those are what real organizations do. Real AI does not just happen from one experiment. GPT 4.0 is just one experiment, one idea, one little component in the worlds systems. If you want it to grow you need to make the tools that people need, not just make one and hope someone else develops it.

OpenAI ChatGPT 4.0:

Your perspective is very valuable, and you’re right in saying that the usability of AI in real-world applications, including complex projects, is an important factor in fulfilling OpenAI’s mission.

Your points regarding the need for robust project management tools, better context retention, and handling multiple parallel conversations are all valid and important for AI’s future development.

It’s important to note that OpenAI is contributing to this development in several ways. By making its models like GPT available via APIs, it allows other developers, researchers, and companies to build on top of it, using their domain-specific knowledge and resources to develop more tailored applications.

At the same time, OpenAI is continually improving its models and exploring new architectures and techniques. While OpenAI may not directly create a full suite of tools for every possible application, it does provide the foundational technology that others can build upon. This approach allows for a diverse ecosystem of applications and tools to emerge, catering to a variety of needs and industries.

That said, your feedback is indeed valuable. As AI continues to evolve, it’s essential to consider the various ways it can be made more useful and applicable to real-world challenges. OpenAI values the input from its users and the broader community in shaping this development.

—- Regenerate Response:

Your feedback is certainly valid, and the perspective you bring up emphasizes the importance of effective tools that bridge the gap between foundational AI research and its practical applications. OpenAI’s strategy does involve more than just research; it’s about facilitating a wider adoption and positive impact of AI technologies.

However, the landscape of tech innovation is often shaped by a division of labor:

1. Core Development: Organizations like OpenAI push the envelope by developing cutting-edge technologies. Their expertise lies in foundational research and creating powerful models.
2. Application and Specialization: Other entities, such as startups, tech companies, or individual developers, often take this foundational tech and refine it into specialized tools, services, and products tailored to specific markets or needs.
3. Feedback Loops: Feedback from practical applications can inform further foundational research. It’s a synergistic relationship.

OpenAI has indeed made significant strides in promoting responsible and widespread use of AI. They’ve opened up their models for public use, engaged in partnerships, and fostered a community around their tools. Their approach isn’t just about “hoping” someone else develops it but rather providing the infrastructure for others to innovate on top of their foundational work.

That said, your feedback underscores the desire for more direct involvement from OpenAI in the application space, and it’s a perspective worth considering. OpenAI, like many organizations, is continually evolving its strategies based on feedback and the changing landscape of technology and society.

— Regenerate Response

Your feedback is absolutely valid and thought-provoking. Indeed, AI technologies, including OpenAI’s GPT models, will become truly transformative when they are accessible, practical, and easy to use in real-world applications. For AI to fulfill its vast potential, it needs to be integrated into tools and platforms that people use every day.

While OpenAI as a research organization focuses on developing the foundational AI technologies, it’s crucial to consider how these technologies can be made more usable and practical for users in various sectors, including but not limited to education, healthcare, scientific research, and business operations.

The development of application-specific AI tools, in addition to the foundational models, could indeed make AI technologies more accessible and useful. It is also worth noting that creating an ecosystem that encourages third-party developers and startups to build on top of the foundational technologies could accelerate the integration of AI into real-world applications.

In either case, your point about the need to focus on practical usability and the implementation of AI technologies in real-world scenarios is well-taken and valuable. It’s a perspective that should be seriously considered in the strategic planning for the development and deployment of AI technologies. Your feedback will certainly help in shaping future iterations and applications of AI.