Palantir and coffee: Models are the Commodity
Vložit
- čas přidán 5. 08. 2024
- This week I break down the All In Pod and their insights into AI and how Palantir will enable the future they've finally glimpsed.
blog.palantir.com/technical-c...
Join the discussion: / discord - Věda a technologie
Just listened to all in episode and I immediately thought about palantir as well. I thought this is exactly my thesis for palantir… they’re starting to get it. But they just haven’t connected the two yet
Great video codestrap. Thanks for all the insight you provide. Much appreciated !
Merry Christmas Code! Awesome to see the future Palantir has been preparing for play out in real time.
Great Pods Codestrap. I was thinking of Palantir when I was Watching All In this Morning. So Glad you brought it up. If you do a little digging you will see David Sacks is already invested in Palantir. Keep up the good work.
I just finished watching the all-in pod and thought of you!!!!
Thanks for another great vid. I look forward to these now.
Thank you so much for everything Codestrap! You are the heart and soul of this community
Thank you!
Merry Christmas! 🎄🎁 Great video and a ton of super good information as always. Thanks for your content.
Thanks!
Thank you for explaining the ai and machine learning
Thanks good sir. I like what you have to say about Palantir. Have a great Christmas!!
Thank you!
I couldn’t agree more on MAAS. I hope to convince some people soon to start trying somethings. Hopefully I will follow up with you next week.
Was literally watching the All in Pod as you posted your vid. Had to switch to your vid tbh 💪🙏
Glad tp see you're also a fan of "THE POD", my guy!
Great video! Many legacy companies treat AI as a magical solution rather than a tool. Which is why they spend lots of $ by putting up an AI engine and throw their massive data sets at it. Instead of focusing on the data they need, how to retrieve it effectively, have a flexible ETL layer and ultimately how to create AI models that align with the data.
Beat me to it. Great video
I think the future of all the applications being predicted in Chamath's podcast + every AI / ML commercial or consumer data solution has the potential to be built on top of Foundry. That would play with the argument that Palantir builds software a decade before they become a necessity. Gotta wait haha
+ As you say, lots of Fortune500 companies have taken notice... Just recently Panasonic (and what it means to its partnerships 👀). Maybe we are wrong, but what I see is; Palantir has already taken a big chunk of this market and is poised to become the standard... 100% conviction! 🧘🏻♂
Thanks for the info, Santa Codestrap!
I watch you on Amit’s channel didn’t realize you had your own you have a new subscriber
Another great video! Thanks for your work! At this, i have to say that i hope your criticisms (e.g. granting access for developers and creating a promoting the creation of a bid developer community) are perceived by PLTR management). I believe there is still a lot of execution risk in that corner.
Agreed
Each time I watched your videos, i buy additional 100 shares 😂
Love the sentiment but remeber this is not investment advice.
LETS GOOOOOOOOOOOOOOOOO
Love the hat!
When I was watching this recent episode of “All In “ today I was thinking of Palantir.What happened to Sachs during this segment? Board meeting ? Really? Sachs who is a buddy of Peter Theil could not be there to mention Palantir when they were basically breaking down where software is going and exactly what Palantir does already.Very frustrating.
I know right. I'd like to hear his take on Palantir within the context of operational GPT and AI generally.
One component I think you guys miss on the SPACs and elsewhere is this: they need to map the ontology of major current verticals AND potential future verticals(SPAC).
How does privacy and data protection impact Palantir's ability to train models on customer data that they can't/won't/don't want to export from the customer for training purposes? Are they able to train it in place on customer data without compromising it and then extract that model for general use in another context/application?
Foundry comes in multi and single tenant stacks. While multi tenant stacks offer software guarantees for tenant isolation, single tenant stacks are deployed inside the customer's network and offer network isolation. Customers can opt for single tenant stacks to provide stronger data protections. Under no circumstances is the customer data ever used by Palantir for any purpose.
Hey code, been a fan of your deep tech coverage and this video nails the point. The piece Im missing here is, when chamath was talking about training vs inference, I can see how the inference models can be deployed almost readily at production/iot/edge environments by apollo. What do you think about training or fine-tuning these models on proprietary data? That needs huge HPC infrastructure investment, where do you think pltr fits in to this part of the equation?
They operationalize the data by providing not only the source connections and software defined data integration technology (which drive data integration costs to effectively zero), but also the platform manage the feedback loops for continuous model training and deployment. Palantir is very unique in this regard by providing version control and CI/CD for not only data pipelines but also ontologies and models.
In other words the orchestration is out if box. I've often described Foundry "infrastructure from code" as opposed to "infrastructure as code".
I see how this resonates with operating system for the entire data layer and serving as a vertically integrated ci pipeline into the HPC infra instead getting fed from a silo NFS and deploys the trained models through CI (apollo) for inference! Truly brilliant.. Thanks for connecting these pieces together.. 🙏
Don't understand anything but it sounds promising.
almost FIRST
First
What’s the release date for the palantir documentary?
I have this exact problem at work. Large large company migrating legacy applications. React/nextJS to get SSR + all the other crap I’m sure you’re aware of. Curious how Apollo plays any role? Can we connect?
Yes, reach out by DM on Discord. Theres a link to the server in the description
CS, join Twitter community! You can contact these guys directly. All-In, $PLTR weekly and your video all in the same weekend - mind blown
Adding PLTR on every down day.
Isn’t it mutualistic not symbiotic relationship?
🎅🏻
Could Palantir’s tech be used for forensic analysis of FTX 🤔
Yes it could. They have a Palantir for Financial Management video on their channel. However FTX is going to be way more of an investigation than an analysis, at least at first. My understanding is they obfuscated their finances.
CodeStrap makes me want to buy 1000 more shares lol
there's no innovators dilemma if you ask the innovation how to mitigate the innovators dilemma 😆
One of my most bullish multi bagger stock. Too bad the market cap isn't still 1b but I have a feeling theres about a 50% chance this baby can 100x from here
just one huge problem. Companies will not entrust their propiertary data to third party models even if you give them legal garentees. The models will have to be built inhouse.
The models will never be the commodity because every company uses different models.
I watched All in podcast for a few weeks and concluding they are all full of nonsense
You are doing a great job on your channel. It could help you to use a site such as "promo sm"!