How Peter Thiel’s Palantir helped expand the NSA’s global spy network

Re:How Peter Thiel’s Palantir helped expand the NSA’s global... by new299 on Wednesday February 22, 2017 @11:25AM (#13706645)

I'm not going to say it's false. But I've love to see citations supporting the notion that technological progress has slowed since the 1970s.

Looking around me I see some evidence:

CCD Cameras: https://en.wikipedia.org/wiki/Charge-coupled_device#History LEDs: https://en.wikipedia.org/wiki/Light-emitting_diode#Initial_c... Laser diodes: https://en.wikipedia.org/wiki/Laser_diode#History LCDs: https://en.wikipedia.org/wiki/Liquid-crystal_display#History Microprocessors: https://en.wikipedia.org/wiki/Microprocessor#History

All invented before 1970. However, they were not really in the hands of consumers by this time. So, I wonder if the point of invention, or the point at which they really had widespread availability is more important?

[ Parent | on HN ]

Re:How Peter Thiel’s Palantir helped expand the NSA’s global... by nickpsecurity on Wednesday February 22, 2017 @11:46AM (#13706854)

Cloud computing is the mainframe model. ALGOL was safe, maintainable language for systems with LISP being ultimate for scripting or flexibility. Current languages mostly do what the could already do. High-performance, NoSQL databases were how mainframes started. Clustering & early distributed systems by Dec's VMS & Tandem NonStop. Xerox PARC came up with desktops. Englebart's people demonstrated all kinds of stuff in 1968:

https://en.wikipedia.org/wiki/The_Mother_of_All_Demos

Most stuff I see today is just rehashes of concepts that were made from the 1960's to 1980's. The core ones you need for personal communication or running a business were all built before 1970. There's still innovation happening but most improvements were just incremental rather than ground-breaking. The most groundbreaking stuff today, like a computer in your pocket, were often result of massive, economic investment into a pile of incremental improvements by many people over time. I'm fine with this but let's not act like we're having many Xerox PARC or Doug Englebart moments.

[ Parent | on HN ]

Re:How Peter Thiel’s Palantir helped expand the NSA’s global... by jdminhbg on Wednesday February 22, 2017 @12:22PM (#13707168)

> Cloud computing is the mainframe model.

It really, really isn't. It shares one characteristic (computation done on behalf of one client by another remote client), but pretty much nothing else. The need to pigeonhole all developments in terms of historical patterns is just revisionism.

[ Parent | on HN ]

Re:How Peter Thiel’s Palantir helped expand the NSA’s global... by trevyn on Wednesday February 22, 2017 @12:56PM (#13707491)

Computation done on behalf of one client by another (more powerful) remote client, with a timesharing model, is kind of the defining characteristic of both, though. ;)

[ Parent | on HN ]

Re:How Peter Thiel’s Palantir helped expand the NSA’s global... by jdminhbg on Wednesday February 22, 2017 @02:11PM (#13708262)

That's not at all the defining characteristic of modern cloud computing; if it were, you would call VR systems (laptop or phone does computation on behalf of video system on your face) "cloud computing."

What's different about cloud computing is the persistent connection and availability, combined with the ownership model and the resulting implied scalability.

[ Parent | on HN ]

Re:How Peter Thiel’s Palantir helped expand the NSA’s global... by acoard on Wednesday February 22, 2017 @03:13PM (#13708766)

>persistent connection and availability

I can't really comment on the ownership model, but surely a persistent connection and availability is an incremental change.

[ Parent | on HN ]

Re:How Peter Thiel’s Palantir helped expand the NSA’s global... by jdminhbg on Wednesday February 22, 2017 @10:47PM (#13711460)

If always-on-from-everywhere is incremental, then what isn't incremental? Are smart phones an incremental change from desktops?

[ Parent | on HN ]

Re:How Peter Thiel’s Palantir helped expand the NSA’s global... by nickpsecurity on Wednesday February 22, 2017 @11:26PM (#13711623)

If we give you that, they have one breakthrough over mainframes. That makes them "mainframes that are always on from everywhere." Mainframe apps were already doing that with terminal, thin clients, web apps, and so on. It was part of modernization efforts to give workers or customers better access to data or business logic on mainframes by putting better interfaces in front of them. Situation is improved today with cloud stuff but core capability was there.

[ Parent | on HN ]

Re:How Peter Thiel’s Palantir helped expand the NSA’s global... by nickpsecurity on Wednesday February 22, 2017 @01:01PM (#13707542)

Mainframes combined most of the muscle in centralized location for weaker, thin clients (eg terminals) to use remotely. Alleged security and efficiency benefits of centralization with one set of staff handling issues for third parties. The CPU, memory, and I/O was metered with many locations charging departments per use. Used VM/370 to virtualize OS's that could run in parallel for various workloads. Scale up quickly. Later clustering and Sysplex for high availability with multiple datacenters. Specialist packages (aka platforms) for various languages or use cases. Hardware accelerators for I/O, databases, crypto, and so on.

Was I just reading a list of benefits and features of modern clouds? Young folks would think so but that was all mainframe stuff from 1960's to 1980's with key hardware & VM tech done by early to mid 70's.

[ Parent | on HN ]