Talent, pay, performance and management.

Anonymouslty contributed as a comment on the Science key to nuclear labs future says Chu post and moved here because it is so interesting:

To conduct business the labs need a cadre of talented and experienced technical staff including material scientists, chemists, weapons phyicists, engineers, technicians and intel types to keep our capability current. Since the timeframe to learn this arcane technology is a long one, incentives are needed for these folks to stick around.

Since the cold war ended, weapons science alone can't attract enough adequate talent; so one needs attractive scientific thrusts (NIF, Fusion, lasers, ACI, HEAF) to interest them.

So on examination it appears that the country needs some above average technical people paid at an above average rate to aid retention.

Now the over-paid managers argument that you raise seemed to have merit. Until I realized that I only noticed management here - in 30 years - when it was bad. And there is plenty that was not so good, including my own.

It actually takes a long time to become an effective technical manager, one that has seen enough both success and mistakes to know what a team must do to accomplish work. Mistakes occur along the way, big and small, benign and significant. A very experienced manager is somewhat invisible to the team and avoids most mistakes. He may add 15% to a top down cost estimate because it was not adequately scoped, or take 15% off of a bottoms-up because scope is double counted. He may even structure a team to work well together in the way the team organizes; reports and works together. He may occasionally see mistakes before they happen. And with long experince he begins to grow and reward his own talent over time, rather than just raiding others (something Moses must still learn).

A good example of recent questionable judgement, possibly due to low salaries attracting middling talent, is when the core UC team managing the labs agreed to bid on the NNSA contract to a manage LANL and LLNL at the transition. We, the country, the Congress, and the Lab may have been better served if the core team had said "NO" when asked to bid on the new contract... rather declining with... "as you have this structured this new contract, it will not work well and we refuse to participate in dismembering the functional entity in which UC invested so much."

Gutsy move, an all of nothing gamble. But the perhaps the only choice with a chance of keeping the Dingell-lead retaliation against Los Alamos mistakes from ruining the Design labs. But I believe the the beginning of the deterioration of LLNL will be marked by historians as the last day of Contract 48.

No $50k kid out of school is going to stand in peril for what's right. He doesn't know what's right and has no investment in the success of the institution, nor the financial independance to be courageous.

GM's founding genius knew this. When Durant consolidated the disparate elements into the operating divisions when GM was founded, he made sure the leaders were very well paid. So well paid they could use independant judgement, to make sure the Division got the best management. LLNL needs this.

Currently LLNL can't get the $300k-per-year folks to do what's right. What we need are managers with talent, experience, independance and iron balls. How much would a Johnny Foster (TRW), a George Schultz (Bechtel) or a Sam Teng (MIT/CERN)take to work here?

This argues instead for above average compensation to retain above average talent in the core scientific staff, with a few geniuses very highly paid to be the bullwark against the undiscovered threat.

And we need some multi-million dollar per year management folks, 'cause what we are paying now ain't attracting talent adequate to surmount the challenges that NNSA places in their path.

Or Chu could drop D'Agostino, whose performance is an argument against low pay for senior management.