NeurIPS 2023 Vendor Hall
8:19

NeurIPS 2023 Vendor Hall

Yannic Kilcher 14.12.2023 10 478 просмотров 273 лайков

Machine-readable: Markdown · JSON API · Site index

Поделиться Telegram VK Бот
Транскрипт Скачать .md
Анализ с AI
Описание видео
Some impressions from the vendor hall of NeurIPS 2023 Links: Homepage: https://ykilcher.com Merch: https://ykilcher.com/merch YouTube: https://www.youtube.com/c/yannickilcher Twitter: https://twitter.com/ykilcher Discord: https://ykilcher.com/discord LinkedIn: https://www.linkedin.com/in/ykilcher If you want to support me, the best thing to do is to share out the content :) If you want to support me financially (completely optional and voluntary, but a lot of people have asked for this): SubscribeStar: https://www.subscribestar.com/yannickilcher Patreon: https://www.patreon.com/yannickilcher Bitcoin (BTC): bc1q49lsw3q325tr58ygf8sudx2dqfguclvngvy2cq Ethereum (ETH): 0x7ad3513E3B8f66799f507Aa7874b1B0eBC7F85e2 Litecoin (LTC): LQW2TRyKYetVC8WjFkhpPhtpbDM4Vw7r9m Monero (XMR): 4ACL8AGrEo5hAir8A9CeVrW8pEauWvnp1WnSDZxW7tziCDLhZAGsgzhRQABDnFy8yuM9fWJDviJPHKRjV4FWt19CJZN9D4n

Оглавление (2 сегментов)

Segment 1 (00:00 - 05:00)

have you thought of this actually like let's say someone comes in here grabs this and walks out what are you going to do well it's a duplicate so there's not actually any of the chips in it so see that's why I asked this is not actual but Nvidia would be upset and charge us probably however okay I see all right sorry about the poor quality live stream we're going to dive into the sponsor boost again and then later we'll do a poster session how about that lab they still have a company but apparently they recognized people look there's look deep mines but there's Demis escaping to us can't be bothered to be here it's just on the screen I know what to do so many Financial companies here here's Market there sit down I think what I struggle with is actually what I can do the most well because and um and Dallas like there's a real Hub of like software engineer Community there and a few finance companies have made some of the finance companies here yeah it's yeah you never see them in like papers and stuff but then they come here and they grab people they hire them you never see the people again look again quantity trading hi how's it going have y great um I'm a brand Parker so I'm the one non Tech person on this bus uh like imp poster you know cool what are you doing here I'm just uh reporting a bit from the conference so why would anyone work here and not at this trading firm or that several cool uh features of our company we're just 4 years old uh but we're really rapidly growing yeah uh we are more than 140 people now we are totally investor free we don't have any external money and we don't accept any external money we are uh trading only on our money from this scratch we have really strong quantitative researchers team and we're using machine learning we don't have any manual trading machine learning was focus on deep learning and serious analysis all the team can code including office managers like we had hackaton for the whole company in non Tech category like it was hackaton for all like for everyone but there were different streams uh for like winners and on text stream third place took office manager from Dubai it was her first time coding in Pyon ever in life but she could do this like the easiest strategy so he did it what you're doing the company as it is there is a pot of money you're coding machine learning code to invest that pot of money into markets with the goal of growing the pot of money yes and also to provide liquidity on the markets yeah yes and also like creating you medicine use of machine learning and de learning and developing this also the field itself not just only for making money because we have a lot of researchers that's really good at the field not just because they're City and trying to earn some money in for St trading it's just one of the fields that can use it yes that's fair so yeah and also we have full merch we have like really nice working culture we have our average age of in place 30 years yeah so we're kind of young people uh full of energy it's really cool to work with some like you know nice uh energized people with some really nice jokes yes yeah a well thank you so much you're welcome here's Lambda are these actual h100s yeah they are so we're looking at 200k or something like this right here even more nice and then we have the GH 200 over there too that's pretty a neat shap and you bring this just to flex I've never seen the heat it's not mine well it's soon mine I guess if you look away a bit you might be able to make it to the street have you thought of this actually like let's say someone comes in here grabs this and walks out what are you going to do well it's a duplicate so there's not actually any of the chips in it so see that's why I asked this is not actual but Nvidia would be upset and charge us probably okay I see but they very cool so you rent these out yeah we do yeah and we sell the physical infrastructure as well so if you want to purchase these to own and like deploy yourself on premise or in our Co location we do that too cool very nice all right there is Astra zenica and they Pres presumably do some

Segment 2 (05:00 - 08:00)

health stuff I guess I'm not really into health I'm a very unhealthy person in general and here is Bosch hi all I know of Bosch is that they make like drills and when I see Bosch AI I imagine like a power drill with like an llm in it and I need to prompt it to start dring is that what you do yeah that's what people we have also Foundation model for dishwasher go drill things step by step and it will like start drilling no yeah that that's the we are getting a lot like do you put the foundation models into the dishwasher so these are the question we are usually getting but I mean uh so we can talk to uh like on a practical example like autonomous driving so autonomous driving is usually based on the sensor like can be camera lar but then you don't really uh you still can't really achieve that fully autonomous driving because it can't drive like a human right like but to achieve like drive like a human you need to go beyond the like the you know those conventional methods yeah and that's where for example we can kind of bring the reasoning capability of let's say large language model Vision language models to actually achieve let's say fully autonomous driving and that's what we are doing like now for example we try to see how we can actually bring let say language model into autonomous driving and this way it's not going to be only captioning like where just kind of commentate about let's say possible action for example if it's making a turn we want to know why it's making that turn also we want actually to improve the like that turn her problem like so that's what actually how we are doing this for our products pretty cool yeah big SN day thank you Apple very fancy very minimal what you expect like an Apple Store learned about from your videos that's like a thing I use a lot so yeah ni oh wowy me I'm a big fan sorry we didn't mean interrup your video there is IBM okay look at this is like St Stars an equal crew here look giant group of people Yan Lon in the but we have no idea how to train robot to share up the video this is something that barel about

Другие видео автора — Yannic Kilcher

Ctrl+V

Экстракт Знаний в Telegram

Экстракты и дистилляты из лучших YouTube-каналов — сразу после публикации.

Подписаться

Дайджест Экстрактов

Лучшие методички за неделю — каждый понедельник