ravishingravi Posted January 28 Share Posted January 28 He got the message. Link to comment Share on other sites More sharing options...
ravishingravi Posted January 28 Share Posted January 28 @bharathh Link to comment Share on other sites More sharing options...
ravishingravi Posted January 28 Share Posted January 28 No coincidence they announced around trump inauguration Link to comment Share on other sites More sharing options...
bharathh Posted January 28 Share Posted January 28 1 hour ago, ravishingravi said: @bharathh AI models are not a bunch of matrices with floating point numbers. Stopped reading there. Link to comment Share on other sites More sharing options...
diga Posted January 28 Share Posted January 28 3 minutes ago, bharathh said: AI models are not a bunch of matrices with floating point numbers. Stopped reading there. decision trees can be modelled that way.. Markov chains Link to comment Share on other sites More sharing options...
bharathh Posted January 28 Share Posted January 28 1 hour ago, diga said: decision trees can be modelled that way.. Markov chains True - but it is not for all AI models - esp Gen AI ones like Deepseek which is what he attempts to explain later in the tweet Link to comment Share on other sites More sharing options...
kepler37b Posted January 29 Share Posted January 29 On 1/26/2025 at 8:48 PM, ravishingravi said: What's ironic is that the so called free society of creativity and excellence didn't release the paper or made open source. These guys have put a detailed paper out there including their failures. Amazing stuff. https://arxiv.org/abs/2405.04434 Hans want glory. The end goal is to proclaim "saaar..we izzzz arrived!!!". Hence this open source circus. Nothing more.....Nothing less. Link to comment Share on other sites More sharing options...
G_B_ Posted January 29 Share Posted January 29 We might have a tech bubble on our hands re valuations... The key thing is 6 million usd. If true then that is the kicker. Is the value added by a 300k usd engineer in the bay area worth it compared to the 150k engineer in London or 100k engineer in tel aviv. Is the 100 billion data centre offering the same gains as a 20 billion data centre in eastern europe. Its not just genetative ai with llm. Isro routinely achives what nasa did so 40 years ago at fraction of the cost. Byd is undercutting tesla. The overarching question is, is the usa about to out price itself out of the data processing industry. Salaries in the usa are simply not globally competitive. randomGuy 1 Link to comment Share on other sites More sharing options...
G_B_ Posted January 29 Share Posted January 29 I also think many in europe who had conceeded this entire industry to the usa will sit up and take notice. It is possible for Europe to compete at those investments despite the regulations. I can see the creation of another entity like airbus. Link to comment Share on other sites More sharing options...
ravishingravi Posted January 29 Share Posted January 29 Link to comment Share on other sites More sharing options...
ravishingravi Posted January 29 Share Posted January 29 4 hours ago, G_B_ said: I also think many in europe who had conceeded this entire industry to the usa will sit up and take notice. It is possible for Europe to compete at those investments despite the regulations. I can see the creation of another entity like airbus. Europe is done. They are not interested in these things anymore. Europe is on secular decline. Best case scenario is reducing cost through immigration and keeping growth flat. MechEng 1 Link to comment Share on other sites More sharing options...
G_B_ Posted January 29 Share Posted January 29 They said that on airbus vs boeing. Fair to say that 6 million is so low, not even regulations will get in the way Can see european companies setting up shop in switzerland. Link to comment Share on other sites More sharing options...
MechEng Posted January 30 Share Posted January 30 On 1/27/2025 at 7:06 PM, bharathh said: To be honest - I feel the next big thing in tech is going to be computing architecture. The one that is able to make quantum computers portable or work on the next gen of architecture to be able to optimize AIs is going to be the big winner. Software will reach a plateau soon without some breakthroughs in architecture. Something to advice the next gen in case someone asks you what they should pursue as a degree. This. Link to comment Share on other sites More sharing options...
MechEng Posted January 30 Share Posted January 30 On 1/28/2025 at 5:38 AM, bharathh said: We have invested a lot of money in our Digital India platforms which are the need of the hour. Sure AI is important as well - but the govt is not going to be running these platforms. Let private players in India step up. The Chinese had nothing to show for the past 2 years. I am sure there are ppl in India also working on these things - just not production ready as yet. Doesn't mean we have nothing. LMAO, no. Link to comment Share on other sites More sharing options...
MechEng Posted January 30 Share Posted January 30 (edited) 35 minutes ago, BlueBlood said: This is the idiotic brainwashed nonsense. Chinese had nothing to show for last 2 years? Where do you illiterates even come from? Check Tencent's github or Alibaba's. Check models like Qwen or PixArt for images. Or even on video side with HunYuan which beats OpenAI Sora in all metrics. Chinese AI universities have been #1 in research papers. TikTok whole algorithm is AI powered not including DJI with their drones. Until when do you fools hide behind the fact that no media in India is honest so you blabber falsehoods like this and getaway as if it's fact. Reality is Ambani and Adani are busy spending hundreds of millions of dollars on weddings and boxing matches like the one coming up with Connor McGregor vs. Logan Paul. They have not contributed a single rupee to IIT or other research labs. So much talent in IIT with zero funds as Modi is too busy charging 28% GST and sucking the money out of middle class and poor people. And SBI and LIC is too busy loaning to Bajaj, Adani and Ambani - same old money. I personally know many IIT PhD's moving to US as there's no compute available to extend their research. Yet idiots here worship the same oligarchs who are glorified rent seekers and not innovators. I've read some research papers where they were attempting to apply game theory to optimize procurement negotiations strategy. It will be insane if they actually come up with a fool proof solution. China is far from perfect, but say what you want - they are real global leaders in research today. Edited January 30 by MechEng coffee_rules 1 Link to comment Share on other sites More sharing options...
diga Posted January 30 Share Posted January 30 11 minutes ago, MechEng said: I've read some research papers where they were attempting to apply game theory to optimize procurement negotiations strategy. It will be insane if they actually come up with a fool proof solution. China is far from perfect, but say what you want - they are real global leaders in research today. Nash equilibrium doesnt address it? Link to comment Share on other sites More sharing options...
MechEng Posted January 30 Share Posted January 30 12 hours ago, ravishingravi said: Europe is done. They are not interested in these things anymore. Europe is on secular decline. Best case scenario is reducing cost through immigration and keeping growth flat. Europe are like the Parsis from 18th century. They showed strong loyalty to Indian kings when they escaped Iran from persecution. When the British came, their loyalty shifted to the empire to adapt with changing times. P.S.: Not maiming Parsis here. ravishingravi 1 Link to comment Share on other sites More sharing options...
MechEng Posted January 30 Share Posted January 30 (edited) 8 minutes ago, diga said: Nash equilibrium doesnt address it? It's not been applied yet in the real world because there are too many boundary conditions to factor in, or else the results will lead to inaccurate answers. However, there are few papers from Chinese universities who are exploring game theory applications to procurement and logistics. Haven't come across a US research paper on the same yet. Edited January 30 by MechEng Link to comment Share on other sites More sharing options...
diga Posted January 30 Share Posted January 30 38 minutes ago, MechEng said: It's not been applied yet in the real world because there are too many boundary conditions to factor in, or else the results will lead to inaccurate answers. However, there are few papers from Chinese universities who are exploring game theory applications to procurement and logistics. Haven't come across a US research paper on the same yet. Search papers on MIMO/xtalk cancellation using DSM from 2006-2014 ... might have a few pointers about non-cooperative games Link to comment Share on other sites More sharing options...
bharathh Posted January 30 Share Posted January 30 1 hour ago, BlueBlood said: This is the idiotic brainwashed nonsense. Chinese had nothing to show for last 2 years? Where do you illiterates even come from? Check Tencent's github or Alibaba's. Check models like Qwen or PixArt for images. Or even on video side with HunYuan which beats OpenAI Sora in all metrics. Chinese AI universities have been #1 in research papers. TikTok whole algorithm is AI powered not including DJI with their drones. Until when do you fools hide behind the fact that no media in India is honest so you blabber falsehoods like this and getaway as if it's fact. Reality is Ambani and Adani are busy spending hundreds of millions of dollars on weddings and boxing matches like the one coming up with Connor McGregor vs. Logan Paul. They have not contributed a single rupee to IIT or other research labs. So much talent in IIT with zero funds as Modi is too busy charging 28% GST and sucking the money out of middle class and poor people. And SBI and LIC is too busy loaning to Bajaj, Adani and Ambani - same old money. I personally know many IIT PhD's moving to US as there's no compute available to extend their research. Yet idiots here worship the same oligarchs who are glorified rent seekers and not innovators. What is brainwashed about what I wrote? There is a reason why - despite all the products you mentioned - Deepseek has made such a splash. 1. It is Open Source. Despite the current models being hosted in China - ppl can choose to create their own hosting for it which prevented anyone from using the other models as they are suspicious of the CCPs control. Even going forward only OSS from China is going to be accepted to the world at large due to the issues everyone has with CCP control over businesses in China. In 2 years since OpenAI came out with ChatGPT 3.5, Chinese products have made great strides - before that they were nowhere. GPT3.5 was a Sputnik moment for them to step their game up which they have. They have the advantage unlike the rest of the world that their products have no alternatives within China - and so have a captive market for them to evolve. Good on them - but other countries will struggle to keep up with OpenAI, Meta etc. as they are cheaper and better quality that anyone can produce. Barrier to entry for other countries to keep up is high. Best would be for India to leverage OSS models and fine-tune them to create our own versions - rather than trying to build something from scratch. Infra costs can be targetted as a potential differentiator compared to others. Look at what Groq is doing. They leverage other ppl's models and provide blazing fast infra at a pittance and leaving the creation of LLMs to others. You cannot compete with a captive market - nor the quality of R&D in the US. Better to find an alternative than try to emulate and fail and waste money. We have always been masters of jugaad. We should leverage this rather than go the foolhardy route of trying to compete for R&D until we have the necessary resources in place. Infra and architecture can be our strength. 2. Digital India products are completely native and the impact they have are awesome. NPCI, AAdhar, ONDC etc. are revolutionizing e-commerce and fintech - which is as important (if not more) than AI. Why do you think most of our products have gravitated there right now? They have a practical impact for the citizens of the country. AI is still in its infancy. Governance, reliability, hallucinations, non-determinism etc. are problems that still do not have good solutions yet. This is preventing AI from becoming standard or defacto in solutions. IMO, a second-mover advantage in AI till these things get sorted out is worth it IMO. Link to comment Share on other sites More sharing options...
Recommended Posts