Anru Stander retweeted
JUST IN: WHO urges passengers who took Airlink flight 4Z132 from St. Helena to Johannesburg on April 25 to contact health authorities over hantavirus concerns.
Anru Stander retweeted
Project Genie is a prototype web app powered by Genie 3, Nano Banana Pro + Gemini that lets you create your own interactive worlds. I’ve been playing around with it a bit and it’s…out of this world:)
Rolling out now for US Ultra subscribers.
Anru Stander retweeted
Erika Kirk forgives the man who murdered her husband:
"My husband Charlie...he wanted to save young men just like the one who took his life...that young man...
I forgive him.
I forgive him because it was what Christ did and is what Charlie would do."
Anru Stander retweeted
The key to life: Show up. Show up when it’s hard. Show up when it’s messy. Show up when no one’s watching. Show up when you don't feel like it. Show up when the rewards are uncertain. Just show up. You can never bet against the person who just keeps showing up.
Anru Stander retweeted
Former CTO of Loom (Sold for nearly $1b) says life is meaningless & bleak after sale...
Another proof point that life with no meaning and purpose is dull at its core.. Regardless of how much money you have.
What do you think?
Anru Stander retweeted
🔮Palantir explained in the most basic way:
People still ask “What does Palantir do?”. Many are disingenuous, but some honestly still don’t understand it. Here’s how I view Palantir at its most basic form:
Let’s start off with the name of the company… “Palantir”. The name comes from the “Palantir” in Lord of the Rings. The Palantir is an orb that allows the user to see into the future.
Much like the Palantir in the movie, the company exists to help users see into the future with their data. Palantir provides real time insights as well.
Palantir has three main products: Foundry (Commercial), Gotham (Government), and Apollo (DevOps automation). Foundry and Gotham both operate with the same concept of plugging in every data source possible to create what is effectively a digital twin; whether it be for a company, hospital, or on the battlefield. At first, that might not sound very exciting. However, if you imagine how large the government space is or how large a company like Coca Cola is, it becomes clear that there are thousands of data sources — most of these sources are not integrated.
DATA INTEGRATION
Palantir can ingest all types of data sources, including images, PDFs, hand-written notes, video, audio, spreadsheets, etc. This provides a single source of truth.
I’ll provide a real life example. I work as a small business loan officer for a relatively small company (around 60 employees). Even with our company being small, I can clearly see where data integration is game changing. For clients, we have loan applications, intake files, financial statements, client information forms, business information forms, records for hours worked with each client, etc. For my company, we have invoices, financial statements, loan loss reserves, data on clients per industry, loan repayment schedules, etc. None of these sources of data are integrated. We have a garbage loan platform that is supposed to be “integrated”. However, there’s no real integration--it more so just acts as a place to dump our files in. No real insights. Since Palantir can ingest all types of data sources, it would be able to create a digital twin of our company. With this digital twin, I would be able to look at ALL of the data in my company through one single window. I imagine how incredible it would be — I would be able to see how all our clients are performing by industry, which clients are starting to see a decrease/increase in profit, which loan officers have the highest success rate on our loan books, how loans are performing by region, which loans might be missing required documentation for auditing compliance, we would have one single SOP for the entire company and each job role; the possibilities are endless.
My personal example is at its most basic form. Again, we are a small company. Now try to imagine this at the scale of a company with thousands of employees and tens of thousands of data sources. Whether the data is stored with AWS, Azure, legacy systems, etc., Palantir can integrate the data… and they can do it fast. Clients have reported getting fully onboarded within weeks sometimes. For other companies, it is a giant headache to accomplish something like this and they most certainly don’t get it done within weeks.
With this data integration, you can build the skeleton of your “digital twin”, as shown below. The data integration is critical for Palantir, but is not the only feature. In fact, far from the only feature.
*basic example of the pipeline builder in Foundry
ONTOLOGY
In order to make the data integration truly actionable, there must be an “ontology”. An ontology by definition is a Greek word that means “Set of concepts and categories in a subject area or domain that shows their properties and the relations between them." In the context of Palantir, it can be described as the nouns and verbs of your business. I have also heard it be described as incorporating the “logic” of your business.
I’ll give another personal example. As I mentioned previously, I am a small business loan officer. There is a strict process. Each employee has their own role in the process. Some roles create a bottleneck and the process can’t be continued until a certain task is fulfilled by a particular role. In this context, the ontology incorporates all of this nuance into the “logic” of our company (logic of our digital twin). See this concept applied to an airline industry below:
HARMONY
With the data integration and the ontology combined, the data is now harmonized. With this, you can now gain real and actionable insights from your data. This can be performed in real time, or with simulation to play out scenarios in the future.
AI MODELS
Straight from Palantir’s website: “Building AI/ML models for one-time insight is relatively straightforward, but managing models over time and making them useful for downstream operations is much more challenging. Palantir makes AI/ML models operational by deploying them on top of a trustworthy data foundation and continuously improving them based on user decisions and feedback — in a cyclic way. Our platforms provide the matrix for end-to-end MLOps, propelling AI/ML out of the experimental vacuum and into the real world.”
“Palantir offers a new way to manage models by attaching each model to a specific organizational objective. Objectives ensure that business logic is tied to specific business KPIs and deployed for a consistent purpose across use cases. Each objective collects business goals, training and evaluation data, and result metrics — as well as the underlying logic.”
As you can see, the data foundation is the framework that allows the AI/ML models to be operational.
EXAMPLES FOR GOTHAM
- Gotham’s targeting offering supports soldiers with an Al-powered kill chain, seamlessly and responsibly integrating target identification and target effector pairing. Operators experience enhanced situational awareness and effectiveness as Gotham streamlines critical decision-making in the modern battlespace.
- Gotham enables the autonomous tasking of sensors, from drones to satellites, based on Al driven rules or manual inputs for human-in-the-loop control. Gotham empowers you to make informed decisions, maximizing the effectiveness of your assets in even the most dynamic operational environments.
-Harness the full power of the platform from operations centers to the edge, transforming any bunker or outpost into an instant command center with mixed reality capabilities. Immerse yourself in a dynamic environment that bridges the gap between physical and digital worlds.
Watch this 5 minute video on how Gotham is used for defense security:
youtube.com/watch?v=rxKghrZU…
EXAMPLES FOR FOUNDRY
- Global solutions firm Jacobs deployed Palantir Foundry at Agua Nueva Water Reclamation Site in Pima County, Arizona, to accelerate the site’s digital transformation and unlock predictive operations across the plant — a proof of concept that could scale across more Jacobs sites, and to their clients. Results at Agua Nueva, an already-optimized plant: 20% plantwide power savings, Reduced electricity usage from ~2,000 KWh per million gallons treated to ~1,500 KWh per million gallons treated, Eliminated operational fines, and Reduced greenhouse gas emissions.
palantir.com/impact/jacobs/
- Global Renewable Energy Producer Sonnedix has deployed Palantir Foundry at their Sonnedix Talayuela solar plant in Spain — and several other solar energy sites — to radically digitize production processes and reduce plant downtime. Impact: 100% remote monitoring of solar trackers, Immediate notification of solar trackers in need of maintenance, Severely reduced idle time between identification of a tracker issue and resolution by needed repairs — as a result of computer vision models, Expectation to reduce revenue losses associated to tracker failures by at least 20%, Expectation to reduce mechanical failures due to trackers being stuck in a vertical position to 0, Reduced carbon emissions
palantir.com/impact/sonnedix…
-Cardinal Health partners with Palantir to securely integrate clinical data and access dynamic purchase decision insights. Foundry's Al/ML capabilities empower health care systems by securely integrating and modeling diagnosis, clinical, and purchasing data, resulting in a clinically-integrated supply chain solution that can respond to changes in real time.
newsroom.cardinalhealth.com/…
- Ferrari uses the Palantir Vehicle Testing System to analyze prototype data from their fleet, and to conduct tests with precision and speed. Hundreds of terabytes of data are ingested into Palantir from different sections of the Ferrari fleet: from sensor data on specific prototype parts, to raw telemetry data after every type of test, to metadata. The software alerts Ferrari to anomalies and outlier events during testing, dramatically decreasing the team’s root-cause investigation time. Using AI and LLMs, engineers can extract and analyze relevant information, map this to vehicle sensor data and prioritize issues for investigation.
palantir.com/impact/ferrari/
APOLLO
“Apollo centrally manages software across independent environments, regardless of their location or level of consistent connectivity. With Apollo's compliance-aware change management engine, organizations are able to automatically orchestrate software upgrades and changes safely across connected and disconnected, air-gapped environments through an end-to-end process we call autonomous deployment. Apollo manages this with built-in controls that are required for strict accreditation frameworks including FedRAMP, IL5, and IL6.
Apollo's goal is to free developers and operators to focus on innovation and value creation by making it simple to deploy software across many environments of different types. With Apollo, developers and operators do not need to spend precious time on building automation or becoming experts on the intricacies of different kinds of infrastructure.”
In other words, Apollo is like the control center for software updates and management. Imagine you have a bunch of different devices, like laptops, desktops, and tablets, each running on Foundry, but they’re all in different locations, across the world. Normally, updating and managing all these systems would be complicated because you'd need to physically access each one or send updates manually. Apollo solves this by automatically and continuously managing and updating software across all these devices, no matter where they are. It's especially useful for companies or governments that run complex, mission-critical systems that need to be always up-to-date and secure. It's like the AI to deploy and continually update Palantir’s products.
See this example from Morgan Stanley, which operates within one of the strictest regulatory environments on earth: youtube.com/watch?v=E6t_SzQU…
AIP
AIP stands for “Artificial Intelligence Platform”. I view this as the platform that brings everything together. With AIP, you can use natural language to build with Foundry/Gotham. AIP also uses AI to help suggest ways to structure/clean your data. I believe Palantir’s AIP is evolving into a platform that anybody can use. The “generate” function allows you to describe what you’re trying to do, and then AIP builds accordingly. This is a very important aspect of AIP that I believe will be critical for tapping into the SMB market in the future. I imagine a future where anybody, regardless of IT experience, will be able to build with AIP through the generate function… “text to build”.
- AI-driven data pipelining: Integrate LLMs into your pipelines and run them at scale—with error handling, automatic retries, guaranteed outputs schemas, and other production-caliber tooling included.
- AI-driven logic building: Build, test, and release feature-rich, AI-powered functions that leverage the Ontology within your applications.
- Workflow visualization: Automatically visualize the objects, functions, and actions that connect your ontology to your applications. Understand the provenance between these entities.
- Workflow monitoring: Log actions taken and track Ontology metrics to improve and debug workflows.
AIP is making Foundry accessible for everyone. In my opinion, AIP is the most exciting development to come out of Palantir in the last couple of years and is a game changer.
CONCLUSION:
Palantir is not just a “data analytics” company, but rather the foundation that enables users to create a digital twin to operate data analytics and AI on top of. Palantir provides a full suite of integration and application tools to transform your business / government. When people think of AI, they typically think of LLMs such as ChatGPT. While LLMs are exciting, they aren’t very useful for governments and enterprises without the logic of the entity incorporated. This applies to many other AI tools as well. Palantir makes it possible.
This is just the tip of the iceberg, but I hope it helps provide a basic framework for understanding.
Anru Stander retweeted
We recently chatted to Andries Levi Pretorius, from @volleystudent, who shared that high-demand in this sector is why he is so passionate about their developments in Pretoria.
Check out Volley Baileys here👇
platform.easyproperties.co.z…
#EasyProperties | #EasyEquities
Anru Stander retweeted
Many people think Christianity is:
I do whatever | want & however | want. And, God exists to help me accomplish my dreams & goals for my life.
However, Biblical Christianity is:
Jesus saves me from my self. And, I exist to do whatever God calls me to for His Kingdom & glory.
Anru Stander retweeted
𝐓𝐡𝐞 𝐟𝐢𝐫𝐬𝐭 𝐀𝐟𝐫𝐢𝐜𝐚𝐧 𝐌𝐚𝐧 𝐭𝐨 𝐰𝐢𝐧 𝐭𝐡𝐞 𝟐𝟎𝟎𝐦 𝐎𝐥𝐲𝐦𝐩𝐢𝐜 𝐭𝐢𝐭𝐥𝐞 🇧🇼🔥
Letsile Tebogo made that look easy 🤯💥
📺 Stream #Paris2024 live: tinyurl.com/4wrtdwtb
#CloserToYourChampions
Anru Stander retweeted
Friendly reminder: stop building more features at your startup and start marketing the ones you already have.
It’s easy to think more features will bring more users but usually what you have is already enough. The real challenge is letting people know about them.
Highlight what makes your existing features great. Use customer stories to show their value. Talk about the benefits and the problems you solve. Don’t use a bunch of buzzwords and tech jargon.
As Seth Godin said, "Marketing is no longer about the stuff that you make, but about the stories you tell."
Remember less is more and a few well marketed use cases are more valuable than a bunch of underused ones. So please take a break from development and ramp up your marketing.
Anru Stander retweeted
The reason you have an MTN in Africa today is not because a poor boy from South Africa got VC funding and became a billionaire. I know some others try to force a similar narrative but it is not true.
People gained experience, put together plans, and then got investors. No magic. Each base station was built one site acquisition and equipment deployment at a time. Financing was tough but they hired the best to get it.
You just have to work or create work. It is in the creating work that we end up building the things that likely become venture scale. Scale builds upon scale and we have very few things already at scale. Scale is not built upon dreams and wishes but upon solid ground first.
I remember my observations the first time I visited America. My cofounder Salil previously told me that America runs like a well oiled machine. Femi Edun also told me that they build scale by dumbing things down to the basics. I saw that both descriptions were true. The infrastructure exists to sell at scale and you dumb things down to sell more.
According to Ogilvy “we sell or else.” America always sells or else it is not America anymore. To be able to sell, there has to be a market to sell to. They created that market in many ways. Infrastructure, education, and advertising.
You can’t sell when you haven’t built. Prospecting for gold in California led to the great railroads. Mining solid minerals in Congo and Nigeria have only led to wars and terrorism.
In Africa, we build, or else…
Let us build and then sell. We can’t sell what we haven’t built yet. Building on cloud infrastructure far away and trying to provide services to those who don’t have access isn’t building.
The day it dawned on me that payments in Africa wasn’t a startup opportunity but an infrastructure play was the day the scales fell from my eyes. PayPal got licenses in 50 American states and not 50 countries before it started making a lot of money.
One African country will not deliver our version of PayPal without our version of EBay which can only happen because we have our version of UPS or FedEx and the logistics that support it. That is further built on universities and a prosperous middle class.
We don’t have to follow the same trajectory but one truth is evident. Scale builds upon scale. Let’s build scale carefully. We have wasted so much time and effort on trying to create magic on a continent where illusions lead to hunger and despair.
Anru Stander retweeted
My take: Apple has realized they don't need to make a whole car to control the entire car software experience... car manufacturers are so bad at software that everyone just puts CarPlay on anyway ¯\_(ツ)_/¯
BREAKING NEWS: Apple cancels the Apple Car project after a decade-long, multi-billion effort to rival Tesla. Some employees shifting to Generative AI teams. bloomberg.com/news/articles/…
Anru Stander retweeted
I would have done this differently though. I would have called it an EE subscription fee (it is what it is) where you can get discount based on your thrive status. Call a spade a spade!! Otherwise people get confused and it feels “hidden”.
This has the Discovery model written all over it. People won’t be happy, but in the end @EasyEquities has to work on generating more revenue and I believe this is a good start. Always knew it will come. Will it work? We’ll have to wait and see. I’m willing to pay the R25 though.
Anru Stander retweeted
Replying to @AnruStander @EasyEquities
I would have been happy if they come forward and just said “Hi guys we will be implementing a subscription fee of R25 and this is how it works but if your part of our Thrive program and your at level 3 we exclude to you”
Anru Stander retweeted
The software engineering behind the X "algorithm" is epic. It has to distill 500 million daily tweets down to a personalized handful it shows each person. This algorithm runs ~57,000 times a second 🤯 ... and each execution takes 220 seconds CPU time, but runs in 1.5 seconds real time on average, which means compute is distributed across 12+ million CPU cores working continuously.
Each step of recommendation algorithm pipeline is fascinating: maintaining real-time interaction graph, embedding spaces for clustering interests, ranking via 48M param neural net, etc. I love that it's mostly been made open source, and hope more companies do the same!































