- Masters of Automation
- Posts
- Are We Missing the Depths Beneath the Iceberg Powering LLMs?
Are We Missing the Depths Beneath the Iceberg Powering LLMs?
Grow Your Perspective Weekly: AI Chip Market, NVIDIA & Adjacent Industries to LLMs.
Alp Uguray
October 22, 2023
⏳ Reading Time: 4min 14sec
Hallucinations aren’t real if you can’t see them; those who understand why will have the keys to power.
The main treasure is at what runs the LLMs: The Realms of Databases, Synthetic data, data-labeling, and the AI Chips Market
💭 Why is this important?
LLMs, powered by the ever-evolving chip industry, have the potential to be mankind's greatest tool. Their ability to process, analyze, and understand vast amounts of data is unparalleled. However, the confluence of economics and technological might, centralized within a few major players, raises questions and prioritizes innovation by sacrificing equity and shared prosperity.
While it's tempting to view this evolution through a dystopian lens, there are signs of hope. Open-source initiatives, governmental regulations, and smaller disruptors in the chip industry might provide the necessary checks and balances. As LLMs become more ingrained in our daily lives, a collaborative approach among industry, academia, and regulators is paramount. The future need not be a zero-sum game.
History reminds us of the power dynamics with every technological revolution. As with the Industrial Age, the entities controlling the means of production (or in this case, computational power) tend to dictate the rules. If the chip market's oligopoly remains unchecked, we may witness an unprecedented concentration of power and influence within the realm of AI.
It’s not hard to imagine that the winners will be Big Tech and some other players who aim to disrupt the industry via building startups.
“These companies already control the largest cluster of AI processors, the best models, the most advanced quantum computing and the overwhelming majority of robotics capacity and IP.”
🌏 How does Generative AI impact the Climate Tech?
Amid the relentless pursuit of superior chip capabilities, we often overlook the silent yet pivotal tangent industries supporting the LLMs. The cooling our behemoth servers is not merely a technical challenge but an environmental imperative. As we accelerate in the domain of AI, it is essential to recognize that the true hallmark of progress will be gauged not just by how advanced our models become, but by the sustainability of the infrastructure that supports them.
As we hail the giants of chip manufacturing, ClimateTech will continue to emerge as a safeguarding our digital and natural ecosystems.
Adjacent industries to Generative AI, starting from Silicon Chips, Databases, Cooling, and ClimateTech will change how we interact with the tech and drive the cost and adoption of LLMs.
Masters of Automation subscribers receive a special invitation to join Demain.
|
The chip market, spearheaded by industry magnates, is in an exhilarating race to not only advance chip capabilities but also to make LLMs more economically viable. As these companies push boundaries, it's projected that the cost to run complex LLM queries will plummet. The democratization of AI? Maybe. But there's a catch. As costs decrease, the centralized power of these tech behemoths could swell.
They hold the keys to AI's kingdom.
💸 Now that is said, here is what is new in the world of AI and automation:
OpenAI Has Quietly Changed Its ‘Core Values’
OpenAI quietly revised all of the “Core Values” listed on its website in recent weeks, emphasizing the development of AGI. It described AGI as “highly autonomous systems that outperform humans at most economically valuable work.” The previous values resemble those of a research lab, while the new ones are more similar to startup jargon.
Meta’s Oversight Board, an independent body that issues binding policies for the tech company, is working on its first deepfake case. It has reportedly agreed to review a decision made by Facebook to leave up a manipulated video of President Joe Biden. Meta said that the video was not removed because it was not generated by AI nor did it feature manipulated speech.
“The Board selected this case to assess whether Meta’s policies adequately cover altered videos that could mislead people into believing politicians have taken actions, outside of speech, that they have not,” wrote the board in a blog post.Nvidia currently controls a majority of the global microchip market, holding an estimated 80 percent, making it a dominant player in AI chip production, including those used in LLMs (source). The firm's microchips and semiconductors are highly demanded due to their effectiveness in powering LLMs and other AI technologies (source). Interestingly, while Nvidia continues to enjoy significant market share and bullish investor interest (source), the chip market landscape might be poised for some shifts.
A key player to watch is Advanced Micro Devices (AMD). Primarily due to its consumer-centric focus, AMD has seen a hit on its stock value. However, the company seems to be leveraging a shift towards the AI chip market, most notably with its development of a new microchip, the MI300, designed for data centers (source). With these advancements, AMD aims to clinch a larger slice of the AI chip market against Nvidia.
Tying into our previous newsletter, Microsoft presents a groundbreaking framework tailored for developing LLM applications. Now, you can have multiple agents conversing, collaborating, and even inviting human interactions to decipher intricate tasks.
Why You Should Care: Imagine a world where one AI agent handles your primary queries and another jumps in for expert troubleshooting! This is the magic of multi-agent collaboration. They streamline conversations, pass vital information, and ensure a heightened efficiency in problem-solving.
Noteworthy Features:
Ease of Development: Dive into LLM applications without breaking a sweat! AutoGen makes it simpler than ever.
Dynamic Conversations: Whether it's the agent's autonomy, the number of participants, or the conversation flow, AutoGen has got you covered.
Inference API: A superior replacement to openai.Completion and openai.ChatCompletion. Expect enhanced tuning, caching, and superior error management.
Up next in our series: The AI-powered revolution in Synthetic Biology. Explore its transformative impact, from food production to advancing longevity.
Reply