Barron Stone

Dan Wallach of Defense Advanced Research Projects Agency (DARPA) wants to move all software written in C to 🦀 Rust 🦀 The TRACTOR program (Translating All C to Rust) is an effort to create approaches to convert C applications to Rust applications. 🎉 This is a big step forward. 🎉 Rust’s memory management model eliminates a whole host of security vulnerabilities (and bugs) that are rooted in dangling pointers, buffer overflows, use-after-free occurrences, double free, and memory leaks; to name a few. This all happens without a garbage collector 🗑️ How does Rust do this? By restricting/discouraging you from directly allocate memory. Rust uses its Borrow Checker feature to enforce these strict set of constraints, and in turn, Rust will allocate & deallocte memory on your behalf by tracking when it is in & out of scope. Read More about TRACTOR ->

8

3 Comments

Future advanced 'starliners' will look more like these (https://bit.ly/49aB40a) We also foresee NOT only our development costs being lower, ultimately Price-per-Seat will be lower as well and provided on much safer form of spacecraft. #futurespace #newspace #newspaceeconomy #newtravel #spaceliner #CTPSpace

2

Finally received the certificate! I am very grateful to RTX Raytheon continual investment into its employees development. One step closer to my goal of becoming an AI/ML Leader! #AI, #ML, #leadership, Artificial Intelligence

49

2 Comments

Fun Usecase for AI: Whiteboard Capture / Transcribing This past week I did a lot of whiteboarding work and, instead of spending a ton of time trying to capture everything then and there, I just whipped out my phone and snapped photos. Today, I loaded those photos into ChatGPT and asked it to transcribe them. The results were solid but not perfect. 1. My handwriting sucks 2. There are natural interpretation errors. Were I to do this again, I'd write neater and do a bit more logical diagraming on the board. Overall, it did a pretty solid job of the basic rote entry and I merely had to clean it after the fact.

9

AI BUBBLE BLUES: This is an excellent post by Dr. Jeffrey Funk about the AI bubble that's inevitably coming. As he wrote succinctly, "When the bubble bursts, it will be a very large pop. If the generative-AI bubble is different than the dot-com bubble, it is most likely for the worse." The comparison with the dot-com bubble is instructive. At the time, the investment community expected a big return quickly. When that didn't happen, it went bust. Only a few, like Amazon, eBay, and a couple of others, survived. Decades later, the promise of e-commerce and the internet has been realized. It just took longer than the so-called "patient money" was willing to wait. Also, dot-coms were very much limited by slow bandwidth at the time. Simple easy to use websites were more successful back then. The "fancy" websites took too long for the pages to open, and the potential customers got frustrated and gave up (me included). In addition, buying things on the internet was a new concept back then and many users were still wary of it. Generative-AI will probably produce great transformations in the decade or two ahead. It will ultimately make money, just not soon enough for today's speculative investors. "Pop" goes the bubble!

9

1 Comment

Did you miss Reindustrialize 2024? That’s ok, we’re just getting started 🌊 Something magical happened in Detroit this week and it wouldn’t have been possible without Mike & Kate “Software-enabled hardware will create more value than SaaS” “Hardware isn’t just cool again, it’s moving at the speed of software” “The future of America is being built by the people in this room” There’s a movement happening Old companies creating new ideas & New companies innovating old ideas Titans of industry are putting their weight behind America and steering towards the future Founders are bending steel around software and creating new capabilities never possible before Investors are funding these ventures with a fresh perspective on value creation Old roads are being repaved and new trails are being blazed by a new class of companies We’re rebuilding the arsenal of democracy with the primacy of winning There will be more to come, much more Hope to see you at the next one! This is how we win 💪🏻🇺🇸 #American #Industry #Innovation

194

14 Comments

BAD BUBBLES: AI, AM, Tulips, and Real Estate. As discussed in the interview below, AI could be the next big bubble to burst, taking down hyperinflated stock market valuations. It’s not the first, nor will it be the last, of the big bubbles bursting. Why? Greed. From “Tulip Mania” in the 1600s, the “dot com bubble” of 2000, and the real estate bubble of 2007, all bubbles are fueled by massive greed. When Gordon Gecko famously stated “Greed, for lack of a better word, is good,” he was wrong. Today, we have three bubbles, the AI bubble, the residential real estate bubble, and the additive manufacturing (AM) bubble. All three are overdue for a massive correction. All three are artificially inflated. In materials technology, the Additive Manufacturing bubble, also known as 3D printing, has been sucking up massive amounts of capital, to the detriment of other worthy technologies. There’s nothing wrong with 3D printing, other than the hype. It’s a tool and nothing more. It’s actually a very good tool for some types of materials, challenging for others, and impossible for certain classes of materials. I’ve been approached many times regarding ceramic matrix composites (CMCs) and 3D printing. Ironically, 3D printing isn’t amenable to composites with 3D continuous fiber architectures (like most of my technologies). The problem is that investment is usually a zero-sum game.  In short, great materials aren’t getting investment dollars because they can’t be made by 3D printing approaches. This is a missed opportunity. All bubbles are bad. They suck up massive capital while they are being inflated. They wipe out massive amounts of capital when they burst. Many people and businesses are hurt by bubbles, even when they have nothing directly to do with the bubble itself. Collateral damage. Bubbles bad!

12

1 Comment

As I begin my doctoral studies into technical debt, I'm plagued by questions about how we in business perceive technical debt. So, I've built a survey to measure the varying perceptions of technical debt. I promise the survey is very short and quick. Please distribute/repost this survey FAR and WIDE so that I may gather as much data as possible to begin my research into this challenging problem! https://cvent.me/MVAOqw Thank you all! #technicaldebt #technologydebt

25

2 Comments

Several folks have heard me express the believe that leaning more into data science should be foundational for systems engineers because of the need to examine things from many different angles. In current practice, we are more like intelligence agents. In that I mean, we identify, develop, and continuously come back to assets in specific corners of a project to build up our insight about what is happening and what problems are on the way. We also trade what we’ve gathered to these folks in order to keep the contacts going.

16

5 Comments

👋🏽 Let’s talk Shadow IT and why it SUCKS. This is a perfect example. I saw it in the Defense Industrial Base and academia all the time because nobody wants to say “no” to the talented engineers making things (aka $$) happen for the company. 😐 We also observed old servers still on networks with no purpose and one company didn’t even know WHERE the retired server was to shut it down and remove it. APT owned the machine and was using it to pivot inside the network. Shadow IT can bring your whole enterprise down. That said, we’ve all been there. It’s super tempting to install whatever apps or hardware you think will help get your work done faster. It gets the job done and it’s often a tool with which we’re familiar or it’s touted as THE solution to your problem. Or we used it in college or others in the community use it all the time. There are lots of reasons really users want what they want. 🎯 Problem is, that can make life pretty hard for the IT and/or cybersecurity folks. These tools often lack proper security, can increase the company’s attack surface and the cybersecurity team doesn’t even know they’re installed. Cybersecurity departments don’t know these tools exist on the network, so they can’t protect them. That means no security patches, no updates, and no backups, which could lead to lost data or potential security breaches. 📕 Even worse, sensitive data could end up exposed, putting the company at risk for legal trouble and compliance headaches. Plus, if something goes wrong, IT is stuck trying to fix tools they didn’t approve in the first place—YOUR problem just became THEIR problem. Okay so how do we deal with it? How do we shut down shadow IT without stifling productivity? Start by making people aware of the risks and encouraging them to use approved tools if possible. Make it as easy as possible to submit tools for review or provide secure alternatives. Regular check-ins or audits can help uncover hidden shadow IT. We should be doing these anyway! The key is finding that balance between keeping things secure and letting people do their best work. #cybersecurity #ciso #informationsecurity #auditing

3

1 Comment

I shouldn't be surprised by the growth in GPT applications and tools but I am - every. single. week. I am astonished at how accessible its becoming. Everything code is edging closer to LCNC (low-code, no-code) reality. Coding itself is becoming even more of a team sport with these tools. If you are not experimenting with tools like this in your daily work, let me make this easy. Just try it with the links below and a bit of youtube. I've been working on my own local and cloud deployments of a GPT that will learn from docs that I give it (AKA a "RAG" LLM) using Python and open source libraries. Its fun just for the learning and there are so many opportunities to apply to business problems. You want to download and configure your own large language model, keeping your data confidential for business or personal reasons? Choose your own adventure - apps like LM Studio let you configure compatible models from the HuggingFace repo and test them all. This - from the comfort a pre-packaged UI that is highly accessible at first glance. Nomic's GPT4all will let you do this too. I'm sure there are multiple others. If you know one, please share for your colleague's sakes. https://lmstudio.ai/ https://lnkd.in/g6wZDEKd

11

1 Comment

I need to get smart on these modular reactors. Anyone have experience with these?

1

4 Comments

It's not everyday that I get to share that we've won a $125M contract: https://lnkd.in/g66VACNA I was deeply involved with this, with many trips to Houston and many discussions amongst company leadership for the last couple of years as we built Gravitics towards this achievement. And I was the Mission Architect for Blue Origin's Orbital Reef. So I've spent years studying space stations and how to make them work technically and financially. A few thoughts: - Space stations are awesome. And expensive. - Cargo is, by far, the most expensive part of the equation. Space stations could be given away for free and most would still struggle to make a profit because cargo can be as much as $90K per kg. What are you selling up there that's worth $91K per kg? - We've been working hard at Gravitics to get that cargo price down. Way down. It's (super) exciting to be charging forward at one of the key barriers to a healthy, human-based space economy. - Axiom is great. I've often been impressed with their leadership. I think the relationship will continue to be wonderful and we'll continue to build strength as we tick through the challenges together. It's an exciting time for space.

99

21 Comments

AI is the new technology frontier and innovation pioneers are creating new pathways and concepts for future development. Has the U.S. government positioned our U.S. based firms to accelerate and outpace foreign development? Or have we handcuffed ourselves to policies that lack vision and hold firm to a risk adverse cultural mindset?

2

I was working some math problems and found GPT deteriorate with its correct answers before I had to remind myself, these are language tools and not artificial brains. They are heavily influenced by context (what you provide in the prompt, what the ChatGPT application layer may pull from the web, etc). Too much unrelated or incorrect context often results in even more hallucinations and user-induced bias than normal. Try the following: 1) Disable the "memory" feature. 2) Always start with a new session. 3) Don't change direction/function/project in the middle of a session. 4) Make sure every part of your prompt is objective, unbiased, and factual. 5) Always provide clear, specific instructions and the validated data you need for your task. #operations #ai #chatgpt #machinelearning #seta #systemsengineering #systemsengineer #usmc #marines #marinecorps

2

1 Comment

Artificial intelligence and nuclear power. I'm not talking about the plot of Terminator 2. It takes more than ten times the amount of energy to run a ChatGPT search than it does a traditional Google search. Makes you wonder how much energy was used in the recent DiPSS proposal submissions... Artificial intelligence requires massive amounts of electricity to run and the big developers are struggling to get to net-zero emissions by 2030. Microsoft is turning to nuclear power. Google is investing in geothermal. As big defense leans into the use of AI, the broader industrial infrastructure will need to develop around it to continue to support it's usage. Maybe next time, instead of asking ChatGPT to come up with a recipe for your frozen broccoli and that dusty can if chickpeas in your pantry you'll save a few watts and grab Grandma's old copy of "The Joy of Cooking". Link to a quick podcast on NPR in the comments.

24

2 Comments

Candid feedback is the best. Integration is a shift of pace from Paradox and there are some deep topics to grapple with. Space, without the crutch of instantaneous speed and hyper populated regions is VERY HARD to pack full of adventure. The best part about this feedback is how it's already churning my brain for how to adjust for book 3 and 4. Always learning!

5

1 Comment

When I say i am a "certified Space, Cyber, and AI/ML Expert," I mean it. What certifies me? Among other things, my Ph.D. in Astrophysics from the University of California, Irvine. Yes, there is a difference between educational degrees and certifications. But I would make the argument that a Ph.D. level education goes way more in depth than any other level of degree program, as it terminates in the requirement of writing a dissertation containing original, publishable research. The Ph.D. student is also left to design, plan, and execute his or her research project from the start with almost no help from the thesis advisor. That is why I would very much make the claim that the PhD is singular among academic degrees in that it very much combines the aspects of a post-secondary education and a professional certification.

2