0 Members and 2 Guests are viewing this topic.
Palantir Technologies, Palo Alto, California, is being awarded a $27,640,000 fixed-price blanket purchase agreement under the Department of Defense (DoD) Enterprise Software Initiative to provide commercial-off-the-shelf hardware, software and services for DoD, the Intelligence community and the Coast Guard. This one-year agreement includes four, one-year option periods, which if exercised, would bring the potential value of this agreement to an estimated $143,800,000 million. The ordering period of the base agreement will be from July 12, 2019, through July 11, 2020. If all options are exercised, the ordering period will extend through July 11, 2024. No funds will be obligated at the time of award. Funds will be obligated at the delivery order level using operations and maintenance (DoD) funds. This agreement was non-competitively procured with a brand name justification in accordance with Federal Acquisition Regulation 8.405-6 via a limited source solicitation and publication on the General Services Administration eBuy web site. Naval Information Warfare Center, Pacific, is the contracting activity (N66001-19-A-0044).
Some of the new hires have been top salespeople at major American companies like Oracle Corp. and International Business Machines Corp. And additional hiring is underway: more than five dozen global positions with “business development” or “leverage” in their titles are now listed on the company’s website. Those positions, said several people, often perform functions that are typically handled by sales teams, despite the absence of "sales" in their titles and job descriptions.The product that most of the new employees are selling is called Foundry, data analytics software aimed at large corporate clients. Current Foundry customers include Morgan Stanley, which uses it to deter insider trading; Merck KGaA, which uses it to speed drug discovery; and Fiat Chrysler Automobiles NV, which uses it to identify faults in its cars before they cause problems.Foundry, and its predecessor Metropolis, performs the same type of big data analysis Palantir’s Gotham does for government groups -- helping users sift through troves of data to mine for patterns and specific information. Compared to Metropolis, though, Foundry requires less customization and far fewer engineers, resulting in a potentially lower cost to customers and fatter margins for Palantir.Foundry took years to develop and is now one of just two products Palantir offers. The other is Gotham, used by federal, state and local agencies and nonprofits around the world. But by pouring resources into Foundry, the company has signaled a change of priorities, moving beyond a focus on government work, toward more seriously pursuing the larger and potentially more lucrative corporate market.
Thiel has been using his power primarily to help staff the Trump administration, vetting candidates to lead the Federal Trade Commission and helping elevate associates from his investment firms to serve in the Department of Commerce, the Pentagon, and even the National Security Council. Another Thiel ally, Yale computer scientist David Gelernter, also won an audience with the president to potentially serve as his science adviser. Back in San Francisco, Thiel’s employees have reportedly begun referring to their boss as “the shadow president.”
More worrisome than Thiel’s influence over staffing are his unusual views on the relationship between science and society. While the presence of a technologist in Trump’s inner circle should be welcome, Thiel’s brand of techno-futurism often tends toward the apocalyptic, with the billionaire advocating less regulation and oversight to allow experimental developments to advance unrestrained. Like many in Silicon Valley, Thiel subscribes to the idea that one must “move fast and break things” to move forward, and worry about the consequences later. Technological progress, he argues in his book Zero to One, has stalled in part because the federal government has grown too big, and entitlement programs too generous. He has encouraged promising college students to drop out of school to found start-ups and is a leading proponent of seasteading: building floating libertarian islands in international waters that would serve as experiments in life free of government control. Thiel also has a well-documented obsession with life-extension technologies, including extending his own lifespan with blood transfusions from young people. On Bloomberg TV in 2014, Thiel explained that he was taking human-growth hormone pills as part of his plan to live 120 years. “It helps maintain muscle mass, so you’re much less likely to get bone injuries, arthritis,” he said.At the same time, Thiel is grounded enough to know who in Trump’s administration to befriend to move closer to the levers of power. “Thiel is immensely powerful within the administration through his connection to Jared [Kushner],” a senior Trump campaign aide told Politico. As much as Thiel professes to despise politics, he clearly knows how to play the game.
Palantir Gotham is a platform for “needle-in-haystack” analysis. It's used primarily by government agencies that look for bad actors hiding in complex networks: terrorist cells, trafficking rings, money laundering schemes, vectors of foodborne illness, and so forth. These organizations use Gotham to bring their data sources and systems together, map the data to a common model, and analyze it in one place.When my friends and family ask me to explain Gotham, I tell them to think about your typical TV detective drama — there's inevitably a scene where they've put everything they know on a bulletin board and connected it with pins and string. Now imagine a digital version of that where you can also be really rigorous about the sources of each piece of information, when the information was last updated, and tightly control who has access to that information — all while expanding collaboration on that bulletin board across the organization. We also get the benefit of rich audit signal and use it to detect any sort of misuse (more on that here).
Foundry started as our attempt to create scalable process and rigor around data integration. There's a longer story to tell here, but the short version is that we started out with off-the-shelf orchestration systems for running jobs on a schedule (think Jenkins and Rundeck — things that were more robust than cron). At some level, that worked, and we had to figure out how to scale it, leading to a combination of HDFS, Rundeck/Jenkins, a git repo, and a common language for mutating data.But this was a precarious setup. It didn't give us the flexibility we needed to answer questions about where information came from (how did data and code correspond?). It was hard to run jobs in the right order, and easy to overwrite data. But it worked. And it was appealing to our customers, who saw our engineers using this toolkit they'd built to bootstrap our own data integration, and started asking if their engineers could use it.We knew we had built something valuable, with a very broad market. But there were some fundamental issues we needed to fix. Deprecating Metropolis let us double down on solving these problems. We focused on:Versioning. Foundry explicitly tracks future state, independent of (and in addition to) past state. You can branch out to apply different versions of code against the same chunk of data and track, for each version of the data, which version of the code was used to create it. So you can understand what you knew at a point in time, and how the data has evolved since.Branching. Building a more explicit orchestration system, and cleaned up the general idea of the “pipeline.” Instead of a system that just moves data from point A to point G, we built a system that lets you move data from point A to point G, then look back at point F and say “Hey, that was interesting. Let's try some different, random variation, but make sure A-G is still happening.” Work is safe by default, and you have the freedom to test novel ideas without impacting other users.Truly “democratizing” data. Creating a front end that empowers a very broad range of users to engage with data. We wanted people to be able to explore and adapt all the data they could access, in ways that are typically limited to very technical users. Today, Foundry is a platform that provides universal, secure access to all of an organization's content, for decision makers at every level, from the factory floor to the executive office.Another thing we realized as our commercial work expanded was that data suffers from diseconomies of scale: the more there is, the harder it is to actually use it. So as companies generate more data, they have an increasingly hard time using it on demand. And this problem compounds across roles and functions. These companies are complex in many ways — technology landscapes, org structures, shifting strategies, etc. — and they need to cut through that complexity to be nimble in a changing world.With that in mind, we built Foundry to create data-driven loops: use data to make a decision, then use data to assess its impact, in lockstep with all of your colleagues and collaborators. Foundry maintains provenance and attribution so people can trust what they discover and learn, and meets users where they are so everyone can bring data into their daily work. Imagine a manufacturing giant where data scientists, quality analysts, assembly line workers, and executives use data as their lingua franca. That level of collaboration was a galvanizing force in building out Foundry.Collaboration is also why it was so critical to get the version control piece right — we knew that was key to engendering trust in data. It was also critical that we get security right, and fortunately we had a lot of knowledge to draw on from building Gotham. Many of the factors that complicate broadening access to data in the government are at play in multinational corporations, too. One of Gotham's key components is a form of Attribute-Based Access Control (ABAC). Gotham does this at huge scale, with very complex attribute relationships, so we knew how to think conceptually about implementing this for Foundry as well.
Palantir specializes in data-gathering and analysis, most of which it does for government agencies. It has about $1.5 billion in federal government contracts alone, including, recently, with the Space Force and the Navy. Now, as new Covid-19 case numbers break records daily, Palantir is trying to help organize the information with a new platform called HHS Protect, which will be run by another private company called TeleTracking. This partnership has effectively replaced the Centers for Disease Control and Prevention’s (CDC) National Healthcare Safety Network, per the Trump administration’s orders to hospitals to stop reporting their information to it. HHS Protect, which is not accessible to the general public, is now the only source for this information.“Today, the CDC still has at least a week lag in reporting hospital data,” Michael Caputo, assistant secretary of the HHS for public affairs, told the New York Times. “America requires it in real time. The new, faster, and complete data system is what our nation needs to defeat the coronavirus.”Palantir, the architect of this complete data system, isn’t a household name like its Palo Alto peers, but the 17-year-old company founded by Peter Thiel is one of the most valuable private companies in Silicon Valley. That anonymity is a feature, not a bug: Palantir does most of its work for the government, including national security and intelligence operations. In recent years, headlines about the company have stressed its access to everything about all of us, which privacy advocates have long criticized. Palantir’s data-mining software has been credited with killing Osama bin Laden (a claim that has never been confirmed) and blamed for tearing undocumented immigrant families apart.Now, the notoriously secretive surveillance startup that the White House is entrusting with the nation’s coronavirus data is about to go public.
Palantir’s work with various police departments across the country has also brought renewed scrutiny to the company, especially in light of recent protests against police brutality. Palantir’s software powers the Los Angeles Police Department’s predictive crime program, called Operation LASER, which tries to identify and target potential criminals for increased surveillance. The program ended in 2019 amid doubts that predictive policing was an effective crime deterrent, as well as criticism from civil rights organizations that it unfairly targeted minority communities. It’s hard to get exact numbers on how many police departments Palantir has contracts with, but New Orleans’s and New York’s police departments are known customers, and Palantir boasts on its website of its work with the Salt Lake City police department.
