TZUTC: -0500   
   MSGID: 33.consprcy@1:2320/105 2c0227f7   
   PID: Synchronet 3.20a-Linux master/acc19483f Apr 26 202 GCC 12.2.0   
   TID: SBBSecho 3.20-Linux master/acc19483f Apr 26 2024 23:04 GCC 12.2.0   
   BBSID: CAPCITY2   
   CHRS: ASCII 1   
   DeepSeek and the race to surpass human intelligence   
      
   Date:   
   Fri, 31 Jan 2025 07:45:25 +0000   
      
   Description:   
   DeepSeek and the race to surpass human intelligence   
      
   FULL STORY   
   ======================================================================   
      
   Back in October, I met with a young German start-up CEO who had integrated    
   the open-source approach by DeepSeek into his Mind-Verse platform and made it   
   comply with German data privacy (DSGVO) standards. Since then, many rumors   
   have been circulating that China has chosen a different architectural   
   structure for its foundation modelone that relies not only on open source,    
   but is also much more efficient, requiring neither the same level of training   
   data nor the same compute resources.    
      
   When it comes to DeepSeek, this is not a singular breakthrough moment.    
   Rather, AI development continues on an exponential trajectory: progress is   
   becoming faster, its impact broader, and with increasing investment and more   
   engineers involved, fundamental breakthroughs in engineering and architecture   
   are just beginning. Contrary to some market spokespeople, investors, and even   
   certain foundation model pioneers, this is not solely about throwing infinite   
   compute at the problem; we are still far from understanding core aspects of   
   reasoning, consciousness, and the operating model (or software layers) of the   
   human mind.    
      
   Additionally DeepSeek is (was) not a government-sponsored initiative;   
   supposedly, even the prime minister was surprised and visited Hangzhou to   
   understand what was happening. Although Scale AI founder Alexander Wang    
   claims that China already has a significant number of powerful H100 GPUs   
   (about 50,000), yetbased on U.S. export lawsthis fact is not publicly   
   acknowledged. DeepSeek is reported to have only about 150 engineers, each   
   earning in the range of $70100k, which is eight to ten times lower than top   
   engineering salaries in Silicon Valley.    
      
   So, regardless of whether they have powerful GPUs or whether $6 million or   
   $150 million was invested, it is nowhere near the billionsor tens of   
   billionspoured into other major AI competitors. This example shows that   
   different engineering and architectural approaches do exist and may be    
   waiting to be uncovered. Most likely, this is not the ultimate approach, but   
   it also challenges the current VC narrative that its all about compute and   
   scale. Moreover, the open-source mindset behind DeepSeek challenges the   
   typical approach to LLMs and highlights both the advantages and the potential   
   risks.    
      
   Sam Altman is rumored to be hosting a behind-closed-doors meeting with the   
   Trump administration on January 30th, where he plans to present so-called   
   PhD-level AI agentsor super agentic AI . How super this will be remains   
   unclear, and it is unlikely there will be any public declaration of achieving   
   AGI. Still, when Mark Zuckerberg suggests Meta will soon publish substantial   
   progress, and Elon Musk hints at new breakthroughs with Groc, DeepSeek is    
   just another breakthrough that illustrates how fast the market is moving.    
      
   Once agentic AIs come online, they introduce a structural shift: agentic AI    
   is not about merely responding to a prompt, but about pursuing a goal .   
   Through a network of super agents, massive amounts of data are gathered and   
   analyzed, while real products and tasks are delivered autonomously. What is   
   interesting about Sam Altman not making a public appearance and release, his   
   meeting with the U.S. The government hints at potential risks and   
   consequences.   
      
   We are at the Verge of Hyper-Efficiency and Hyper-Innovation    
      
   What we are seeing is the compound effect of investment and ever-growing    
   teams working on these models, with few signs of a slowdown. Needless to say,   
   any quantum breakthroughs would be the next frontieressentially AI on   
   steroidswhere the magnitude of change could increase exponentially. On the   
   positive side, this can unleash innovations in health and medicine like never   
   before in human history.    
      
   In the near future, broader access to AI tools will probably benefit   
   infrastructure providers and hyperscalers such as AWS. It is unclear if this   
   will put NVIDIA at a disadvantage or actually benefit it: as everyone joins   
   the AI race, there could be more demand for compute, not just from big U.S.   
   tech players like OpenAI. Meanwhile, Anthropic and OpenAI run closed   
   ecosystems, but DeepSeeks public paper shares many of its core methods.    
      
   The greatest risk to the U.S. and its current AI dominance is that China does   
   have talent and the strong work ethic to keep pushing forward. Trade    
   sanctions wont stop that. As more engineers come together and keep working,   
   the odds of major breakthroughs increase.   
      
   The Battle of Distrust    
      
   Globally, the U.S. is losing trust. The dont trust China narrative is fading   
   in many parts of the world. While Donald Trump on the surface gains respect,   
   global leaders are quietly looking for alternatives in the background to   
   mitigate. Europe and other Asian nations dont want to be hostage to U.S.   
   technology and will open up to new options.    
      
   Technology doesnt evolve overnight, and weve only seen the start of the   
   breakthroughs to be announced by Groc, Meta, and OpenAI. Simultaneously, new   
   capital will continue pouring in, and other regions will join the race, now   
   that its clear money alone isnt everything. The future might not necessarily   
   be bad for NVIDIA, either, since data centers could appear everywhere,   
   enabling a more global roll-out of AI and creating opportunities for many.   
      
   From Prompting to Action    
      
   There are still numerous smaller AI companies that have received massive   
   funding purely on hope and hype. Yet new approaches to foundation modelsvia   
   architectural and engineering innovationcan continue to drive progress. And   
   once we hack biology or chemistry with AI, we may see entirely new levels of   
   breakthroughs.    
      
   Looking toward the rest of 2025, we can expect more super-agent    
   breakthroughs, as agentic AI and LQMs (Large Quantitative Models) push   
   generative AI beyond fun language-based tools to genuine human worker   
   replacements . Not only will financial modeling and analysis be optimized,    
   but also executionthe entire cycle of booking, planning, and organizingcould   
   shift to autonomous agents. Over time, these integrated, adaptive agents will   
   replace more and more use cases where humans currently remain in the loop.   
   This might also be one of the biggest threats to society: coping with extreme   
   pressures on market economies under hyper-efficiency and hyper-innovation. In   
   2025, we are likely to see breakthroughs in education, science, health,   
   consulting, and finance. With multiple compounding effects in play, well   
   likely experience hyper-efficiency and widespread growth.    
      
   However, the looming threats are real. Agentic, at-scale AI can still fall   
   victim to hallucinations, and now anyone with a few million dollars can build   
   their own modelpotentially for malicious use. While a global, open approach    
   to AI can be positive, many engineering and research challenges remain   
   unsolved, leaving high risks. With the U.S. laser-focused on AI, the race to   
   surpass human-level intelligence is on.    
      
    This article was produced as part of TechRadarPro's Expert Insights channel   
   where we feature the best and brightest minds in the technology industry   
   today. The views expressed here are those of the author and are not   
   necessarily those of TechRadarPro or Future plc. If you are interested in   
   contributing find out more here:   
   https://www.techradar.com/news/submit-your-story-to-techradar-pro   
      
   ======================================================================   
   Link to news story:   
   https://www.techradar.com/pro/deepseek-and-the-race-to-surpass-human-intellige   
   nce   
      
   $$   
   --- SBBSecho 3.20-Linux   
    * Origin: capitolcityonline.net * Telnet/SSH:2022/HTTP (1:2320/105)   
   SEEN-BY: 105/81 106/201 128/187 129/305 153/7715 154/110 218/700 226/30   
   SEEN-BY: 227/114 229/110 111 114 206 300 307 317 400 426 428 470 664   
   SEEN-BY: 229/700 705 266/512 291/111 320/219 322/757 342/200 396/45   
   SEEN-BY: 460/58 712/848 902/26 2320/0 105 3634/12 5075/35   
   PATH: 2320/105 229/426