It's an old problem
"A distributed system is one in which the failure of a computer you didn't even know existed can render your own computer unusable."
Leslie Lamport, 28 May 1987
77 publicly visible posts • joined 11 Nov 2008
Collecting information from humans has always been an expensive and error prone process. Information should be collected from sensors and systems, and these need to be carefully tested and calibrated.
Surveys in general have become particularly problematic, since people no longer answer landlines choked with spam, and they screen smartphone calls against their contact lists. Polling organizations have been having trouble predicting election outcomes as their ability to get good samples of opinion decreases.
Note that the BLS household employment survey has a 90% significance at the +-600,000 job level and the employer survey has a 90% significance at the +-136,000 job level.
https://www.bls.gov/web/empsit/ces_cps_trends.htm#concepts
AGI that has been trained on information scraped from the Internet is unlikely to result in great economic impact beyond enhancing search into something like "search and summarize". There is simply too much nonsense in the training base.
For economic impact AI will need to be trained on the proprietary internal information base of each major corporation or organization. For example, Intel's AI would have to be trained on everything Intel knows about designing and manufacturing integrated circuits in order for the AI to find a path out of Intel's current mess. Same with Boeing.
None seem to get the correct answer, which is "Do you mean the sound "R" or the letter "R"?
Some of the other test questions are equally ambiguous, so what is being tested is whether the model makes the same assumptions about what is being asked as the average naive English speaking fleshie.
So like buying a $150 refurbished desktop from Amazon or Best Buy and running Linux on it?
$150 seems to be the price for the latest pre Windows 11 capable desktops.
I've been thinking about doing that to replace my 9 year old laptop. Anyone have experience doing so?
Of course, I could probably get much of the same performance gain by replacing the hard drive in the laptop with an SSD.
Since almost all calls are from subscribers with unlimited calling and texting plans, AT&T has no business reason to save much of this data. Even in those limited instances where it is used for billing, it need not be stored many months.
However, the government no doubt requires AT&T to keep it in order to support law enforcement and national security investigations.
>>and stops driving itself into the back of stopped fire engines and such at speed.
These would seem to be the very negative examples needed to train the AI what not to do...
10 million crash records are what is likely needed in order to avoid crashes, not 10 million examples of uneventful trips.
Microsoft Job Cuts Hit HoloLens Unit After Setback on Army Goggles
Microsoft won’t be getting more orders for its combat goggles anytime soon after Congress earlier this month rejected the US Army’s request for $400 million to buy as many as 6,900 of them in the current fiscal year.
https://www.bloomberg.com/news/articles/2023-01-18/microsoft-scales-back-hololens-business-after-setback-on-us-army-goggles
I wonder whether Southwest, with their point-to-point route structure, depends more on the spare capacity of the other hub and spoke airlines to move their crews from home to base and base to flight? In the widespread storm conditions, the other airlines have no spare seats for Southwest crews traveling on standby.
The $90 B is capitalized and becomes equity on which a return has to be made or debt which must be amortized and on which interest must be paid.
So the spectrum license fees are in effect a hidden tax which raises prices to the cellular users.
Had IBM not been forced by DoJ to divest itself of Service Bureau Corporation, and had not the FCC in Computer Inquiry I & II forced the Bell System not to provide integrated data processing services, computing would have always been done by dumbish terminals connected to time-sharing servers, and the detour through on-premises computing would have been avoided.
It sounds a lot like a pitch for more university and academic research funding.
Compared with the US, more Chinese basic research is probably done in government labs and in government controlled industrial labs with universities more dedicated to turning out the manpower and staff them. I'm sure that their academics would like funding to be more like the US, where we fund a myriad of small grants to a multitude of academic principle investigators who actually produce very little.
Once upon a time printers were one of MS's competitive advantages. In the early days of word processing, one of MS Word's few advantages was the length of the list of printers supported. The complexity of PC brands, interface boards, and peripheral makers and models eventually became a "competitive moat" for Microsoft.
There are a number of languages which moved down and right when flipping between the Q1 and Q3 graphs, i.e. they declined in Stack Overflow and increased in Github rank. So maybe rather than a story of no movement in overall rank, there is a story regarding Stack Overflow versus Github. Possibly these languages have other support communities which makes Stack Overflow less important to their users?
Good point. At these rates, it appears to be about equal to the taxes that would have been paid by the manufacturers if they had produced the laptops in the United States. So it obviates the tax advantages of manufacturing offshore.
Plus, it is paid on the value at the border, so it isn't actually 25% (or whatever) of the retail price as the MSM reporting would leave you to believe. Depending on the product and how it is distributed, it is more in the range of half that at retail, given the distribution, marketing, retailing, and general and administrative expenses loaded onto the price within the US.
It was US politicians that encouraged businesses to move production to China for geopolitical purposes to counter Russia and to nobble the Asian Tigers who had become too successful and uppity. Of course we had previously moved business to the Asian Tigers after the Japanese had become too successful and uppity. Now the geopolitical mandate is to move businesses to the rest of SE Asia and South Asia, since the Chinese are too successful and uppity.
Failures of these sorts are vital since they cause providers to exercise their recovery procedures and they cause users to exercise their mitigation, fallback and recovery procedures. Absent randomly occurring failures at some reasonable frequency society would build itself up for real catastrophic failures.
Back in the day of circuit switched digital telephony, the US used mu-Law non-linear Pulse Code Modulation and Europe used A-law non-linear PCM. I was told by a US member of the CCITT (now ITU) standards group that the US had offered to agree to and change to match the Europeans. He was told that regardless of what the US would agree to, Europe would be different.
Softbank owns 80% of Sprint, a US mobile carrier. Sprint is now the smallest, weakest and least profitable of the 4 major carriers, and is thought to be circling the drain. Softbank is attempting to get a merger with T-Mobile approved by the US Department of Justice in order to salvage its investment.
Softbank also owns ARM.
I would think that most females and slightly built Asians and East Africans should not pilot MAXs. Pilots should resemble NFL defensive linemen.
A backup system of mechanically operating the tail surfaces may have been OK on the early, small 737s, but that is a strategy that fails on later, bigger models unless you impose severe physical strength requirements on the pilots.
They have Kylin, a version of Linux, and they have COS for mobiles. These find application in the government and military for obvious reasons.
For widespread use, there needs to be a commercial reason to more widely deploy them. This move by the US may be the trigger that is needed.
Apps, essentially special purpose client software running on the phone, were needed with early generations of mobile data, since they could conserve bandwidth while providing a rich user experience.
With 5G, bandwidth would no longer seem to be a problem, and most functionality could be delivered with a modern mobile browser.
The other purpose of apps seems to be to nickle and dime the user for this and that. However, most useful services can be had for free.
Finally, how many apps does the average smartphone user need? Clearly not the hundreds of thousands that are available in the stores. The vast majority must be downloaded only by the creator, his family and friends. Probably the top thousand apps account for almost all of app usage.