[FRIAM] I am accepting wagers

[uǝlƃ ↙↙↙]

It does. Many of us have had hands-on experience building the kind of tool Biden proposed. The trick is that we, as pattern recognition devices, tend to think our experiences are generalizable. Whoever suggested finding the right 15 people committed that sin. Your doubt is healthy. But while doubt is a good servant, it's a bad master.

It's entirely plausible that a couple of kids could put this thing together, well below $10m. But it's also plausible it'll take multiple millions to do it well. I don't know. I'll have more data after they announce who's doing the work and what requirements they intend to satisfy. Until then, anecdotes serve nobody but the story-tellers' egos.


On 3/15/21 1:00 PM, Frank Wimberly wrote:
> I'm not sure this adds anything.
> 
> As I've said before I was a one of the first people to be hired by Bell Labs to work specifically on No. 5 ESS.  During my interview they told me they wanted to hire 125 computer science PhDs.  The year I finished there were 108 who graduated in the entire country.  Although my dissertation was in numerical methods my first assignment was in software administration methodology.
> 
> As the project grew every day consisted of two 3-hour meetings in the morning and two in the afternoon.  These focussed mainly on the interfaces between the various subsystems.  This was the first ESS to be developed in C and with multiple processors.
> 
> Several times I heard comments like "if we could get the right 15 people in a room together we could get this done."  After a year and a half of this I decided to head for academia.
> 
> If you read about 5ESS on Wikipedia you will learn that it is widely deployed in telephone networks worldwide including cell networks.  It has 100,000,000 lines of C code and 100,000,000 lines of header files.  I'm skeptical about the "15 people".

-- 
↙↙↙ uǝlƃ