|
Pryor on Fire posted:I mean I hope people understand how risk works in here aaahhahahahaha
|
# ? Dec 28, 2016 02:42 |
|
|
# ? Jun 8, 2024 06:55 |
|
FSLR up 4.20%
|
# ? Dec 28, 2016 02:48 |
|
Pryor on Fire posted:I mean I hope people understand how risk works in here, but then again everyone is just posting their % returns instead of their sharpe ratios so maybe not.
|
# ? Dec 28, 2016 03:42 |
|
Josh Lyman posted:As someone who has taught Sharpe ratios to business students, I really hope it's not actually used in industry.
|
# ? Dec 28, 2016 04:36 |
|
Goon project to write a book called "Day Trading For Retirement: The Guaranteed Way"? Maybe we could even market it as "What financial regulators don't want you to know!"
|
# ? Dec 28, 2016 05:09 |
|
Solice Kirsk posted:Goon project to write a book called "Day Trading For Retirement: The Guaranteed Way"? Maybe we could even market it as "What financial regulators don't want you to know!" Just figure out when the government is going to be amenable to keeping the markets happy and load up on beta. Judging on trumps advisors, things look pretty good. But maybe that's just what they want me to think before they implement policy that will collapse everything. Or you can do what I do and buy high/sell low. Not broke yet because I limit my gambling expenses.
|
# ? Dec 28, 2016 06:14 |
|
Josh Lyman posted:As someone who has taught Sharpe ratios to business students, I really hope it's not actually used in industry.
|
# ? Dec 28, 2016 09:18 |
|
https://twitter.com/HansNoordsij/status/813848411611025408 This is pretty cool. I don't trust tesla will make it but we're definitely going to stop driving soon
|
# ? Dec 28, 2016 11:49 |
|
When I visited my parents over Christmas my dad and I spent some time watching both Nvidia's promotional videos and their Physx Flex demonstrations. What's interesting is that Nvidia isn't quite the next Intel - they are sort of the next Intel and Cray Research combined. Crays weren't just huge computers - they operated differently than regular computers of their time doing things like vector-based processing to streamline certain math operations on huge datasets. However back in the 80s and 90s Cray supercomputers that did cutting-edge research across tons of fields were phenomenally expensive I think these performance computing systems from Nvidia cost in the five-figure range. Also I want to dig deeper and see what tools they are developing for both their gaming markets (like libraries and documentation for both game developers and people making deep learning and AI applications.) I want to see more examples of actual solutions made by third parties using Nvidia's systems. Three-Phase fucked around with this message at 12:43 on Dec 28, 2016 |
# ? Dec 28, 2016 12:40 |
|
Quote != edit Check this out - demo of Nvidia's self driving car: https://youtu.be/-96BEoXJMs0 Three-Phase fucked around with this message at 12:46 on Dec 28, 2016 |
# ? Dec 28, 2016 12:41 |
|
Three-Phase posted:Quote != edit If everything that's in this video is true... Holy. poo poo. If one of the big cloud services suppliers start using GPUs instead of CPUs for all their AI related workloads, Nvidia makes a shitton of money. It's a huge investment area right now, too. orange sky fucked around with this message at 13:54 on Dec 28, 2016 |
# ? Dec 28, 2016 13:43 |
Three-Phase posted:Quote != edit This is a good visual demonstration of how machine learning works, now imagine the same methods applied to every job in every industry. It works exactly the same way. People are just starting to gently caress around with this on the side and are getting results like this coming from zero experience in just a few months. In every industry. Pryor on Fire fucked around with this message at 14:12 on Dec 28, 2016 |
|
# ? Dec 28, 2016 14:06 |
|
Pryor on Fire posted:This is a good visual demonstration of how machine learning works, now imagine the same methods applied to every job in every industry. It works exactly the same way. People are just starting to gently caress around with this on the side and are getting results like this coming from zero experience in just a few months. In every industry. Ahhhhh we're all gonna lose our jobs. E: I don't know if I should sell soon though, people will probably sell it in January and cause a panic, but I believe in everything the company is involved in.
|
# ? Dec 28, 2016 14:18 |
It's gonna go up and down and there's a lot of competition off in the distance but yeah this is an "electricity" moment, the poles are starting to go up and everything is going to change.
|
|
# ? Dec 28, 2016 14:26 |
|
Here's another industry: https://twitter.com/daniel_bilar/status/813019450551242752
|
# ? Dec 28, 2016 14:29 |
|
orange sky posted:Here's another industry: Assuming these claims aren't overly optimistic or hiding costs, that's really cool just from a tech perspective. And a creepy future infrastructure aesthetic perspective.
|
# ? Dec 28, 2016 14:39 |
I hadn't even thought about that, my god yes please bring me demented ML designed buildings and bridges that look loving insane. This cyberpunk future is gonna own.
|
|
# ? Dec 28, 2016 14:47 |
|
orange sky posted:Here's another industry: That's got nothing to do with machine learning. It's just optimization. The original story was about 3D printing. http://www.arup.com/news/2015_05_may/11_may_3d_makeover_for_hyper-efficient_metalwork Even then, the part on the left likely costs a fraction as much to produce.
|
# ? Dec 28, 2016 14:50 |
|
Three-Phase posted:Quote != edit I honestly am amazed at the world I'm seeing take shape around me. By the time I retire I fully expect flying robot taxis and hopefully moon vacations.
|
# ? Dec 28, 2016 15:12 |
|
Pryor on Fire posted:I hadn't even thought about that, my god yes please bring me demented ML designed buildings and bridges that look loving insane. This cyberpunk future is gonna own. Imagine the loving spaceships ML will design
|
# ? Dec 28, 2016 15:15 |
|
orange sky posted:If everything that's in this video is true... Google, who seems to be really far ahead in AI right now, is making their own AI-tailored (edit: tailored for AI, not by AI) chips instead, which are way way way better than GPUs or CPUs for those tasks. http://www.tomshardware.com/news/google-tensor-processing-unit-machine-learning,31834.html https://cloudplatform.googleblog.com/2016/05/Google-supercharges-machine-learning-tasks-with-custom-chip.html If those claims are true (and I have no reason to doubt then), then I mainly see one of two things happening: 1) Google becomes the cloud service provider for ML/AI since everyone else's performance is significantly worse. 2) Other cloud service providers follow suit and use custom chips. GPUs don't seem like they'll come into play. Dr. Eldarion fucked around with this message at 15:37 on Dec 28, 2016 |
# ? Dec 28, 2016 15:34 |
|
Dr. Eldarion posted:Google, who seems to be really far ahead in AI right now, is making their own AI-tailored (edit: tailored for AI, not by AI) chips instead, which are way way way better than GPUs or CPUs for those tasks. what about nvidea/intel/amd making custom ML/AI hardware as opposed to repurposing graphics hardware thereby capturing a bunch of value from other cloud providers while reducing costs of service?
|
# ? Dec 28, 2016 15:43 |
|
Why shouldn't I buy VSLR? Stock is trading at 0.5 book value. Sounds like cash flows from existing installs are pretty safe.
|
# ? Dec 28, 2016 15:51 |
|
Aaand it's coming down. I still think it'll hit new highs before the end of the year, every manager wants this in their portfolio.
|
# ? Dec 28, 2016 16:00 |
|
NVDA gapped up and then straight down since open. at least i sold my call at the top. Puts would have been a good idea.
|
# ? Dec 28, 2016 16:05 |
|
Solice Kirsk posted:I honestly am amazed at the world I'm seeing take shape around me. By the time I retire I fully expect flying robot taxis and hopefully moon vacations. Cool post from the 1950s.
|
# ? Dec 28, 2016 16:20 |
|
I'm tellin ya boys once AI gets going the selectrics won't even need a broad to peck at 'em!
|
# ? Dec 28, 2016 16:36 |
|
Lol did the Citron tweet do that to NVDA?
|
# ? Dec 28, 2016 16:54 |
|
As part of my job, I train neural networks (on Nvidia GPUs, using MS CNTK). It is cool stuff but, again, you don't need a fancy GPU to "run" this software - you use the GPU in making it, meaning the audience isn't as huge as I think some people are imagining. We serve a lot of customers using the models we build, but those models are served from boring old web servers; it's only our development environment that has some fancy GPUs. These developments are good for NVidia, but even if the technology becomes very successful and pervasive it might not mean that NVidia explodes. NVidia is quite likely still a good bet, but it's complicated.
|
# ? Dec 28, 2016 17:08 |
|
Might take a big plunge into big old boring CSCO here this year. Still working on the numbers and trade. But if anyone has any thoughts I'd like to hear them.
|
# ? Dec 28, 2016 17:23 |
|
Whats the big advance in ai? Has the hardware just reached a tipping point to enable the 20 year old techniques to do (a lot) more useful poo poo?
|
# ? Dec 28, 2016 17:24 |
|
Whelp...
|
# ? Dec 28, 2016 17:25 |
|
Three-Phase posted:When I visited my parents over Christmas my dad and I spent some time watching both Nvidia's promotional videos and their Physx Flex demonstrations. Fun fact: The Cray X-MP was roughly equivalent to an iPhone 3G
|
# ? Dec 28, 2016 17:37 |
|
Baddog posted:Whats the big advance in ai? Has the hardware just reached a tipping point to enable the 20 year old techniques to do (a lot) more useful poo poo? We've had sort of a confluence of improved software, improved hardware, and some large, confidence-building successes over the last few years. Lots of the stuff currently being done could potentially have been done quite a while ago; it would have been harder, sure, and taken longer, but most of the developments have been incremental. Largely it's just the perception/realization that this stuff is practical - and practical "now" rather than vague R&D - that has opened the floodgates on development.
|
# ? Dec 28, 2016 17:56 |
You can try to point to one thing like machine vision or multi layered neural networks or just the internet now existing, but really it's a convergence of a lot of separate things at once. The NYT magazine just had a good piece about the history of AI and its current state at Alphabetquote:The kinds of jobs taken by automatons will no longer be just repetitive tasks that were once — unfairly, it ought to be emphasized — associated with the supposed lower intelligence of the uneducated classes. We’re not only talking about three and a half million truck drivers who may soon lack careers. We’re talking about inventory managers, economists, financial advisers, real estate agents. What Brain did over nine months is just one example of how quickly a small group at a large company can automate a task nobody ever would have associated with machines. http://www.nytimes.com/2016/12/14/magazine/the-great-ai-awakening.html
|
|
# ? Dec 28, 2016 18:10 |
|
AMD took a dump this morning and it triggered my trailing stop. Still got out a couple hundred bucks profit though.
|
# ? Dec 28, 2016 18:13 |
|
I think a market inflection point is nearing. Everyone is starting to get awfully complacent- consumer confidence is surging higher (and donald trump is sending himself congratulatory tweets about it), the market just keeps marching upwards and everyone seems to think AI is right around the corner and its going to change everything (it will, but its not right around the corner) CNBC was talking about NVDA and how it continues surging up and that 'no one knows what this thing is worth, it could change humanity and everything we know- it could be worth trillions'. Thats good old fashioned bubble talk
|
# ? Dec 28, 2016 18:14 |
|
greasyhands posted:I think a market inflection point is nearing. Everyone is starting to get awfully complacent... I'm not so sure. Depending on what can get through Congress, some coming changes to the tax code could goose a dollar of pre-tax earnings by at least 15%. If that's the case there is a pretty big bullish argument--and the market should have further to go. But that's more of a medium term argument. No idea what's going to happen soon; a correction feels overdue, but it often feels that way.
|
# ? Dec 28, 2016 18:49 |
|
I can't stop laughing at nvda being down 6 points literally one day after someone dumped 20% of their equity portfolio into it
|
# ? Dec 28, 2016 19:14 |
|
|
# ? Jun 8, 2024 06:55 |
|
Agronox posted:Lol did the Citron tweet do that to NVDA? lol yeah, the fuckers. I think they're wrong though Just based on the fact that people have been saying what they said about NVDA since like, $70-$80. I think their 10 series GPUs sold like mad over the holidays + will continue to perform very strongly. Their point #1 just seems completely wrong to me and it's the same general "how much money is there in gaming even " argument being made for years now. Clearly there's a huge market which is growing, and specifically PC gaming is making a decent recovery versus where it was at a few years ago. That's been my rationale for liking NVDA for a while now (if you want to dig up old posts in this thread for example) and the Tesla/AI poo poo is just icing on the cake for me imo. That's not actually where most of their revenue comes from despite most analysts (that I see) getting hyped for AI and robot car stuff. Their GPUs are still selling like hotcakes. Moridin920 fucked around with this message at 19:29 on Dec 28, 2016 |
# ? Dec 28, 2016 19:16 |