Posted March 4, 20241 yr At what point do we start stepping in an putting restriction on it or do we not? Do we treat it with restrictions like how we test bioweapons? More importantly, do any under the radar stocks exist? OpenAI hasn’t confirmed what Q* is, with reinstated Chief Executive Officer Sam Altman only describing it as an "unfortunate leak,” but in the media, it sounds similar to another system Alphabet Inc.’s Google is working on. Gemini is the big new competitor to ChatGPT, which won’t only generate text and images but also excel at planning and strategizing, according to Google DeepMind CEO Demis Hassabis. DeepMind famously created an AI model that beat champion Go players, and Gemini will use some of those techniques for problem solving. With Q*, OpenAI seems to be pushing ChatGPT in a similar direction since, according to multiple reports, Q* can perform grade-school math. That might sound unimpressive, but combining math capabilities with software that can also write text and create imagery is unique, and ChatGPT until now has struggled to do equations correctly. If it could, that might correlate with an improvement in problem-solving. Math requires understanding a problem and figuring out the steps to solve it before carrying out all the right calculations. That process is a little closer to how we humans think and solve problems. https://www.bloomberg.com/opinion/articles/2023-12-04/openai-s-q-is-alarming-for-a-different-reason?embedded-checkout=true
March 4, 20241 yr China and Russia and God knows who else aren’t going to be abiding by any restrictions.
March 4, 20241 yr This might be a stupid question here but one thing that I've wondered about is why some of these semiconductor companies are only focused on making chips significantly smaller... They're using ultraviolet light and lithography and making them smaller and smaller... Instead of printing on small flat wafers why not imprint the transistors into a cube shape material and not worry about size... All these servers are stored in huge data centers anyway... I realize smaller chips means smaller hardware and more complex phones and stuff but why not focus on a more powerful data center that's accessed by remote devices?
March 4, 20241 yr 6 minutes ago, The_Omega said: China and Russia and God knows who else aren’t going to be abiding by any restrictions. China definitely not... They're trying as hard as possible to become leaders in ai tech
March 4, 20241 yr 2 hours ago, DaEagles4Life said: More importantly, do any under the radar stocks exist? To invest in? Everyone and their mother is investing in Nvidia I would think any companies who have vast amounts of data are best positioned to benefit from ai... Facebook Microsoft Google Amazon
March 4, 20241 yr 10 hours ago, Aspiritfall said: This might be a stupid question here but one thing that I've wondered about is why some of these semiconductor companies are only focused on making chips significantly smaller... They're using ultraviolet light and lithography and making them smaller and smaller... Instead of printing on small flat wafers why not imprint the transistors into a cube shape material and not worry about size... All these servers are stored in huge data centers anyway... I realize smaller chips means smaller hardware and more complex phones and stuff but why not focus on a more powerful data center that's accessed by remote devices? Go back to dumb terminal tech? People keep trying to do this, but it makes no sense. The processor in a dumb terminal is then mostly idle, doing nothing, and wasting energy and compute. Smaller circuits allow more computing power in a smaller footprint. Larger requires more resources, usually more power, More powerful data centers are done by clustering servers together into a pool of compute resources. Virtual servers can share CPU, memory, disk & network, making more efficient use of the hardware. Data center space is not cheap. All the built in redundancy, power, Internet, cooling, etc costs money. So smaller servers, using less power is more efficient, especially when carved into resources so that one physical server is running dozens of virtual servers.
March 4, 20241 yr 54 minutes ago, Toastrel said: Go back to dumb terminal tech? People keep trying to do this, but it makes no sense. The processor in a dumb terminal is then mostly idle, doing nothing, and wasting energy and compute. Smaller circuits allow more computing power in a smaller footprint. Larger requires more resources, usually more power, More powerful data centers are done by clustering servers together into a pool of compute resources. Virtual servers can share CPU, memory, disk & network, making more efficient use of the hardware. Data center space is not cheap. All the built in redundancy, power, Internet, cooling, etc costs money. So smaller servers, using less power is more efficient, especially when carved into resources so that one physical server is running dozens of virtual servers. Makes sense when you put it like that I would have thought in house data centers the size would be irrelevant
March 4, 20241 yr 1 hour ago, Aspiritfall said: Makes sense when you put it like that I would have thought in house data centers the size would be irrelevant They rent you a rack with a set number of spaces for equipment. You pay for space, power, and connections, and paying for 99%+ uptime is expensive. Backup batteries, backup generators, dual paths for Internet connections, none of this is cheap.
August 29, 20241 yr On 3/3/2024 at 6:10 PM, DaEagles4Life said: At what point do we start stepping in an putting restriction on it or do we not? Do we treat it with restrictions like how we test bioweapons? More importantly, do any under the radar stocks exist? OpenAI hasn’t confirmed what Q* is, with reinstated Chief Executive Officer Sam Altman only describing it as an "unfortunate leak,” but in the media, it sounds similar to another system Alphabet Inc.’s Google is working on. Gemini is the big new competitor to ChatGPT, which won’t only generate text and images but also excel at planning and strategizing, according to Google DeepMind CEO Demis Hassabis. DeepMind famously created an AI model that beat champion Go players, and Gemini will use some of those techniques for problem solving. With Q*, OpenAI seems to be pushing ChatGPT in a similar direction since, according to multiple reports, Q* can perform grade-school math. That might sound unimpressive, but combining math capabilities with software that can also write text and create imagery is unique, and ChatGPT until now has struggled to do equations correctly. If it could, that might correlate with an improvement in problem-solving. Math requires understanding a problem and figuring out the steps to solve it before carrying out all the right calculations. That process is a little closer to how we humans think and solve problems. https://www.bloomberg.com/opinion/articles/2023-12-04/openai-s-q-is-alarming-for-a-different-reason?embedded-checkout=true Just my two pence... The discussions around advanced AI systems like OpenAI's Q* and Google's Gemini certainly raise important questions about regulation and oversight, as well as the broader implications for technology and investment. Regulation and Oversight Balancing Innovation and Safety: As AI technology advances, there's an ongoing debate about the need for regulation. The challenge is to balance fostering innovation with ensuring safety and ethical considerations. For example, bioweapons are tightly regulated due to their potential for catastrophic harm, and some argue that similarly stringent measures might be necessary for advanced AI systems, particularly those that could have significant societal impacts. Establishing Guidelines: Currently, many experts advocate for developing guidelines and frameworks that focus on transparency, accountability, and ethical use of AI. These might include: Transparency: Requiring companies to disclose how their models work and what data they're trained on. Accountability: Establishing clear lines of responsibility for AI-related decisions and impacts. Ethical Considerations: Ensuring that AI systems are designed and used in ways that align with societal values and human rights. International Cooperation: AI development is a global effort, and international cooperation may be necessary to create and enforce standards that prevent misuse while encouraging beneficial applications. Under-the-Radar Stocks Finding "under-the-radar" stocks related to AI can be challenging but rewarding. These are typically smaller companies or startups that might not be widely known but have the potential for significant growth. Some strategies to uncover these stocks include: Industry Research: Look for companies working on innovative AI technologies or applications that might not yet be mainstream. This can include companies specializing in niche AI applications or new methodologies. Investment Platforms: Use investment platforms and tools that focus on emerging technologies. These platforms often highlight promising stocks and startups in various sectors, including AI. Networking and Conferences: Engaging with industry experts, attending conferences, and participating in AI-focused forums can provide insights into emerging companies and technologies. AI Systems and Their Capabilities Q and Gemini*: If Q* and Gemini are indeed moving towards integrating advanced problem-solving capabilities with text generation and image creation, they represent a significant step forward in AI development. Such systems could transform various fields by combining the strengths of multiple AI functions. Capabilities and Challenges: The ability of AI systems to handle complex tasks like problem-solving and creative generation reflects ongoing advancements. However, ensuring these systems work reliably and ethically is crucial. Issues like mathematical accuracy, understanding context, and generating safe and appropriate responses are all areas that require ongoing research and development. In summary, as AI technology evolves, so too must our approaches to regulation and investment. Balancing innovation with safety and exploring emerging investment opportunities can help navigate the rapidly changing landscape of AI.
August 31, 20241 yr A few months back I was reading about the Army doing some simulation testing with a drone that had an AI targeting system, with the parameter that the drone found a target it needed approval from a human operator. It was going well until the drone kept getting told no, then it targeted the human operator, and then proceeded to go after the targets that it wanted to. So they built in a rule that the human operator couldn't be targeted and ran the simulation again. It was going ok until the drone kept getting told no. This time it didn't target the human operator; it targeted the operator's communications system, and then proceeded to go after the targets that it wanted to. They wanted to create the next great command structure, but all they did was create a super E-4 mafia.
August 31, 20241 yr 3 minutes ago, Bill said: A few months back I was reading about the Army doing some simulation testing with a drone that had an AI targeting system, with the parameter that the drone found a target it needed approval from a human operator. It was going well until the drone kept getting told no, then it targeted the human operator, and then proceeded to go after the targets that it wanted to. So they built in a rule that the human operator couldn't be targeted and ran the simulation again. It was going ok until the drone kept getting told no. This time it didn't target the human operator; it targeted the operator's communications system, and then proceeded to go after the targets that it wanted to. They wanted to create the next great command structure, but all they did was create a super E-4 mafia. it's not a bug... it's feature.
September 1, 20241 yr 10 hours ago, Bill said: A few months back I was reading about the Army doing some simulation testing with a drone that had an AI targeting system, with the parameter that the drone found a target it needed approval from a human operator. It was going well until the drone kept getting told no, then it targeted the human operator, and then proceeded to go after the targets that it wanted to. So they built in a rule that the human operator couldn't be targeted and ran the simulation again. It was going ok until the drone kept getting told no. This time it didn't target the human operator; it targeted the operator's communications system, and then proceeded to go after the targets that it wanted to. They wanted to create the next great command structure, but all they did was create a super E-4 mafia. So basically, they built a mini-Skynet. Awesome.
September 20, 20241 yr Author Bringing back nuclear online! Microsoft deal would reopen Three Mile Island nuclear plant to power AI The owner of the shuttered Pennsylvania plant plans to bring it online by 2028, with the tech giant buying all the power it produces. https://www.washingtonpost.com/business/2024/09/20/microsoft-three-mile-island-nuclear-constellation/
September 20, 20241 yr On 3/3/2024 at 9:13 PM, Aspiritfall said: This might be a stupid question here but one thing that I've wondered about is why some of these semiconductor companies are only focused on making chips significantly smaller... They're using ultraviolet light and lithography and making them smaller and smaller... Instead of printing on small flat wafers why not imprint the transistors into a cube shape material and not worry about size... All these servers are stored in huge data centers anyway... I realize smaller chips means smaller hardware and more complex phones and stuff but why not focus on a more powerful data center that's accessed by remote devices? Size = heat, and higher power requirements. Old computers were slower, generated more heat, and used more power. Size costs money too. Larger computers means larger servers, means larger data centers, with higher power and cooling. It has been found to be cheaper and more power efficient to be smaller. Also faster.
May 21May 21 On 9/20/2024 at 10:04 AM, Toastrel said:Size = heat, and higher power requirements. Old computers were slower, generated more heat, and used more power. Size costs money too. Larger computers means larger servers, means larger data centers, with higher power and cooling. It has been found to be cheaper and more power efficient to be smaller. Also faster.Hi.
Create an account or sign in to comment