Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Strategies & Market Trends : Technical analysis for shorts & longs -- Ignore unavailable to you. Want to Upgrade?

To: E_K_S who wrote (59820)	8/28/2024 10:32:24 PM
From: Johnny Canuck	Respond to of 68310

After hour trading usually does not necessarily reflect traders having absorbed the whole story,especially from the conference call. It is usally pretty thin trading. Some traders will roll over their positions tomorrow. Others will write them off on Friday closer to the close in terms of options. 120 to 121 is the closet support level based of the chart. barchart.com I think I saw someone call $107 on one of the threads. I am not sure we get there except maybe at the open for a brief instance. We will see where the battle ends up. I expect a volatile day regardless. When it gets that volatile intra-day like it has for the last few days some sort of big move is coming down the road for the market. Too many nervous traders open to being set off by any hint of good or bad news. Short term it may be hard to get it to rally. As I mentioned before anyone that waa vaguely interested in the move is in it. Like JP Morgan when the stockboy is bragging about how much money they are making on the stock everyone, "the last sucker", is in the stock.

To: E_K_S who wrote (59820)	8/29/2024 3:40:10 AM
From: Johnny Canuck	1 Recommendation Recommended By E_K_S Read Replies (1) \| Respond to of 68310

First NVDA call I have heard. It is interesting to see how reserved Jenson is as opposed to the persona he has one stage for the rehearsed marketing presentations. A few analysts tried to ask the question a few different ways but there was not a clear answer from what i heard. They still want to know what the Return on Investment (ROI) is for customers. Jensen referred to the fact they are seeing it right away for their own models and renting capacity to other users. The second part of the answer is not good as it implies some companies have excess capacity. Keep in in MSFT and AMZN are intending to rent capacity as a business, small companies in specific verticals not so much. There was a reference to the diversity of their customers in the conference call presentation section but I still did not hear "this is the killer app in each of these sectors". My sense from what I heard and what I know from other presentation is that companies are looking for ways to leverage and use their existing data more effectively. They are still not sure if AI is better than traditional data analytics except that in theory AI should uncover new models, trends and ultimately lead to new ways of using their data. It is still too early to identify killer apps. So no one is making a killing selling or renting their existing models to other clients or at least I did not hear that nor have I seen that yet, There have been attempts to set up market places for that. I have not seen any wide spread traction yet. If there was a killer app that was making in roads in specific verticals I would have heard about it by now. Johnny's speculations, not part of the call: I did not hear anyone say it but it is being implied that everyone is making their own custom models. Other than META I don't know of anyone using embedded AI models are part of their operating models. You see it in some specific divisions and it is hidden for the most part: customer service robots that do the initial intake, robots that monitor your choices and selective try to up sell you, drawing inferences from data sets that would in the past require some one posing a query, a coder coding the question and then getting the answer (structure applications that are cookie cutter but most companies overpaid for). Ignore the fact that hallucinations are a systemic part of AI systems and that fact that if you don't know that the answer should be regenerative AI is at best misleading and at worst will make mission critical mistakes most of what I have seen and heard is a slow evolutionary innovation not the revolutionary innovation AI is being represented as. That said we are at the end of year 2 of the AI tech trend. About 1 year to 1.5 years left to prove themselves as a truly revolution technology or at the very least for competitors to catch up and some aspect to become commodities. Autonomous car still not really a thing. Just testing in jurisdictions that can force existing cars off the road and make the management process easier for the new cars. Drug modelling: you still need the basic research to feed into the models and most of the easy mining of data sets has been done on the data sets that are public or easy to access. Privacy legislation still locks a lot of data away. Robot: While it would be nice to have general purpose robots you can re-program for different task easily. It is still only cost effective in industrial settings. Maintenance to keep the robots in tune so the advance models can do all those new amazing things still limits the practically and cost benefit over purpose built robots with jigs to perform specific tasks. Managing wireless and landline networks: Yes it can reduce the maintenance cost and increase the amount of traffic it can handle, but as seen with the CRWD failure, we are still one update away from a catastrophic failure or in the case of regenerative models one update with bad data incorporate into the model update away. META updates some their models for some aspects of Facebook every 4 hours for example I was told at a presentation. General AI is decades away. The amount of data and processing power to achieve the expectation is exponential in nature. The more AI is used on some fields the less diverse the available data and at some point the models converge to a steady state that limits innovation. Essentially the same answer over and over again. The last parts of just thoughts I have based on the presentation I have seen.