In the early days of streaming services, the ability to choose content was a valuable resource for viewers and there were no reference points as to the relative value of such services, but as streaming services have proliferated over the years, price points have become more visible as battles for hit content continue to push streaming service providers to spend more for content creation, eventually finding that those costs must be passed on to subscribers.  During 2020 and 2021, the COVID pandemic pushed consumers toward almost any kind of streaming content, and monthly bills became secondary to the insatiable need for entertainment, but as the pandemic subsided and subscribers faced the icy fingers of inflation for the first time in many years, they started to think a bit more about what it was they were paying for and how much value it actually had, now that they were no longer tethered to their screens.
As streaming subscribers, we found ourselves in the same situation, subscribed to a number of streaming services, some of which we could not remember why or when we subscribed.  Some of that indulgence was due to billing, as direct billing through Paypal (PYPL) or similar services made monthly charges more abstract and bundling under Roku (ROKU) or other services almost hid them all together.   Eventually the day comes when subscribers sit down and decide that they must ‘figure it out’ and dump those streamers that provide lesser value. 
That said, we were curious to see what the monthly cost of streaming services has done over the last 10+ years, which led us to put together this data on streaming pricing.  The most meaningful timeframe is the period from 2019 to 2023, which includes the year preceding COVID and roughly a year (mid 2022 to current) after the pandemic.  The overall change in price among the major streaming platforms shown here during that period has declined by 10.5%, with HBO (WBD) leading the charge down with a 33.3% decline in aggregate price.  The combined Netflix (NFLX) prices (all tiers) was down 18.8% but Amazon (AMZN) Prime and Hulu were both up over the period by 15.4% and 15.0% respectively, while Disney (DIS) was up 14.3%.  While the composite numbers are helpful, the trends are far more obvious when looking at the charts. 
We do note that the 10.5% composite price decline seem across all categories over the 5 -year period is primarily a result of Netflix’s price decreases (all categories) this year, which puts its pricing below its initial 2007 pricing, which should indicate the increasingly competitive nature of the streaming business.  For those streamers that go back far enough, the 10-year data shows Amazon Prime up the most (81.8%), Hulu up 15% and Netflix up 8.3%, so if you are looking for price stability over time, it would seem Netflix is the choice.  Of course, much in the streaming world is based on content offering, but no matter how you look at content, any evaluation would be subjective, which makes it moot in the ‘Fun with Data’ genre, but Figure 7, which shows top 5 streamer subscribers and y/y growth, indicates that the overall subscriber count continues to grow, albeit a bit down from the 2020 – 2022 years.  Again, that would seem to indicate that subscribers are still shelling out cash for streaming services but streamers themselves are facing a far more competitive world than in early and pre-pandemic years.
We note that the subscriber data, in most cases, does not come from the steaming services themselves, so we consider the data somewhat questionable, particularly that of Amazon Prime, but it does give some visual representation of the trends.

Years ago we saw a mock-up of an OLED device that looked like a large pen.  It had a tab that allowed the user to pull out a flexible OLED display that showed full color images and text, as if you were reading a magazine or newspaper., which you could wirelessly update or switch pages with the press of a button.  As one who has spent many hours on trains commuting to and from work, the idea of such a device remained a futuristic but eminently achievable industry goal in our minds, and we have followed the display space closely ever since. 
It seems that we are getting ever closer to that dream as Samsung Display (pvt) promises to show an updated version of their flexible and rollable displays at the upcoming SID show next week.  The product is called Rollable Flex, and while it is similar to other SDC display products, and those of other display manufacturers, it takes the rollable concept a bit further.  Typically rollable displays are able to expand their surface area by 3x by maintaining a wide circumference around a drum, or in some cases folding across the interior of a device, however the folks at SDC have come up with a display that is able to be more tightly ‘wound’ around a cylinder without damage, and is able to expand its surface area by 5x, with the show demo being 49mm (1.9”) long when rolled to 254.4mm (10.01”) long when unrolled by wrapping it tightly around a cylinder.
While rollable OLED displays exist currently, they do not have the physical characteristics to be tightly wound without showing stress and eventual more serious damage.  While all of the materials in an OLED stack have their own ‘modulus of elasticity’, a fancy way of saying its resistance to being deformed, much of an OLED display’s flexibility is not determined by its substrate but by the material used to create the OLED stack’s anode.  The anode must be transparent if the light from the OLED emitters is to exit the pixel, and the most common material for OLED anodes is ITO, or indium Tin Oxide, which it typically sputtered[1] onto the substrate.  ITO is unusual in that it has both high electrical conductivity and optical transparency, a rare combination, along with ability to be finely etched, however the material is also brittle, the antithesis of what is needed for rollable displays, and is also permeable enough to allow oxygen and water vapor into the OLED stack, which destroys OLED materials.
In rigid OLED displays, the ITO is deposited on glass, and with a second ‘base’ glass substrate, locks the ITO and other OLED materials away from oxygen and water, however rollable displays must be built on flexible substrates which leaves the ITO and other material open to damage.  The solution for this issue is ‘layer’ other impermeable but transparent materials over the ITO, a process called encapsulation, solving the contamination issue.  That said, ITO’s brittleness is still an issue, and we expect SDC has come up with either an ITO substitute or modified ITO mixture that allows the material to have a higher elasticity modulus that exceeds that of current materials, and the odds are that SDC will not reveal the details of the difference from ‘normal’ and less flexible OLED stack components.
While this is all technical, it does pave the way for progressively smaller rollable devices and puts into sight the one-day pocket pen that opens into a large, full color display.  Samsung has already patented a number of ‘hybrid’ devices that use a rolled OLED display and a mechanical pull-out frame that holds the display when open, but these are typically the size of smartphones when closed, still a reach from the pocket-pen newspaper.


[1] Sputtering is a process  that involves the creation of plasma that ionizes the mater5ial (source) which forces the molecules out of the source  and on to a target material as a thin film.

We have both extolled and maligned the recent swath of NLM AI systems put before the public in recent months, and like any new ‘human’ we might meet, we continue to find the good and the bad sides of public AIs.  Our goal is to find instances where we can be both comfortable with the information that is provided to us by the AI, more from an accuracy standpoint and less for the conclusions that the AI might provide, but we are continually disappointed with the results as we have noted previously.  Our latest interaction with Google’s (GOOG) Bard left us with the same feeling of elation and then disappointment as we went a step further than our initial query.
Our initial query was to see if Bard had data relative to the number of TV sets sold in China in 2022.  When we were given a definitive answer, we posed a more complex question about that same data for previous years, and Bard was happy to give us the number of TV set units shipped in China al the way back to 2000.  Once we had that data, we decided to look for broad data that might follow a trend line similar to that of TV set shipments in China, and we started with GDP/capita to see if that tied into TV set shipments.  Bard was also happy to provide that data back to 2000.

​The problem, however, began to become apparent when we asked Bard to show the GNP data against the TV set shipment data.  Bard complied, but when we looked at the comparison, we noticed that the TV set shipment data was different from the data Bard had presented earlier.  We asked Bard about the sources of the original China TV set shipment data and the China Tv set shipment data in the GDP comparison and were told that the data sources were different (Bard told us the actual sources but we will call them source 1 and source 2), and in a number of years, very far apart.  We checked with Bard to see if one source might differ from the other as to units shipped or sales dollars, but Bard said they were both TV units shipped in China for each year.
As we have had a problem with the fact that Bard had previously not specified the differences (shipments or dollars) between other data we asked Bard specifically about why there was a difference between the data sets and we got a set of reasons why the data might be different, such as ‘different methodologies’, ‘different definitions’, different time frames’, and Bias, all sounding like pat answers to such a question regardless of the actual circumstances, so we went a step further and asked Bard why one set of China’s TV set data was chosen for one question and another for the comparison, and here is the unusual answer:
Yes, there was a reason why I chose one set of data for the earlier query and another for the comparison.
For the earlier query, I chose data from the World Bank [GDP] because it is a reliable source of data on economic indicators. The World Bank has a long history of collecting and compiling data on a wide range of topics, including GDP per capita. The data from the World Bank is also up-to-date and reflects the latest trends in the market.
For the comparison, I chose data from [Source 2] because it is a reliable source of data on the TV market. [Source 2] has a long history of collecting and compiling data on the TV market, and its data is used by a wide range of businesses, including manufacturers, retailers, and distributors. The data from [Source 2] is also up-to-date and reflects the latest trends in the market.
It is important to note that both sets of data are from reliable sources. However, the data from the World Bank is more general, while the data from [Source 2] is more specific to the TV market. Therefore, I chose the data from the World Bank for the earlier query because I wanted to get a general overview of the Chinese economy. I chose the data from [Source 2] for the comparison because I wanted to get a more specific look at the Chinese TV market.
It seems Bard did not get the gist of our question, so we asked if the data originally given for China’s TV set shipments was also reliable, but Bard was struggling with the question.  When we were more specific about the data sources from our original question, Bard insisted that the earlier data, despite the fact that it was different from the later data, was still accurate, and cited the same reasons as we noted above as to why the data might be different from different sources.  At that point we realized we would not be able to get any more information as to why one data set was chosen over another in each instance.
We were certainly not trying to find flaws in Bard or any other AI system, in fact we were hoping that we could confirm the data and draw a conclusion as to whether there was any connection between China’s GDP/capita and TV set sales, but the entire exercise points out why NLM’s are not quite ready for prime time, and can do considerable damage if they are not very carefully used and all data is vetted.  Even if a NLM is taught from a specific data set, such as real-time stock market data, or historic product shipment data, the idea that the system would favor different data sources in each instance proves the unreliability of the algorithm.  Any good investor would want to know the data sources for anything involving a major portfolio conclusion and if an analyst were unable to confirm the validity of the data, it could easily invalidate the conclusion.  As data underpins most businesses, without reliable data there is no reason to move from human to AI, at least for now.
 

​Japan has joined the US and its allies in banning the export of a number of semiconductor tools that are used in the production of advanced semiconductors, particularly EUV tools at 10nm – 14nm and lower nodes, and tools used in the etching of stacked memory products.  This follows an agreement with the Netherlands to do the same earlier this month, which choked off all EUV and advanced DUV tool sales from ASML (ASML) to China.  Japan had proposed the rules in March, at which point China had announced some restrictions on Japanese imports, although the Japanese bands are a bit less onerous than the US restrictions.
The Japanese companies most likely to be impacted would be Tokyo Electron (8035.JP), Advantest (6857.JP), Screen Holdings (7735.JP), Nikon (7731.JP), and Lasertec (6920.JP), all to varying degrees depending on product mix, with the Japanese government evaluating potential subsidies or other compensation based on each company’s exposure to the added restrictions.  Based on published data, Tokyo Electron derived 22.5% of sales of semiconductor production equipment from China in fiscal 2023, down from 25.6% in 2022.  Advantest does not break out China specifically but derived 90.6% of 2022 semiconductor & component test (9 months) sales from Asia, with semi test being 71.1% of total sales.  Screen Holdings, for fiscal 2023 (March) generated 80.5% of sales from semiconductor production equipment, and while they do not break out sales by region, ‘overseas’ is 84.3% of SPE sales.  Corporate reps have made little comment, other than saying they will evaluate the situation or China is only a small part of the company’s overall sales strategy.

​The Data Protection Commission in Ireland issued a decision concerning Meta’s (FB) use and transfer of data that was collected in the EU, under its 2018 GDPR (General Data Protection Regulation), that was adopted by the European Data Protection Board.  The decision, the end result of a battle between the EU and Meta over the transfer of customer information from Meta Ireland to Meta US originally led to a judicial review in Ireland back in 2020 based on a draft decision made way back in 2016 over the same issue, which states that Meta ‘failed to guarantee a level of protection to data subjects that is essentially equivalent to that provided by EU law.”
The ruling ordered that the data transfers be suspended and processing and storage of the data in the US also be suspended, along with a €1.2b ($1.33b) fine, to own knowledge the largest fine ever levied under the GDPR.  The original complaint even cites leaks by Edward Snowden, indicating that the NSA was operating surveillance programs on systems operated by some of the largest technology companies, as a basis for the illegality of the transfers to US soil, which tacitly the US confirmed.  More to the point however was the accusation that the data was both collected without the user’s permission, and was used for purposes that user did not agree to
The ruling, which gives  Meta  6 months to halt the transfer practice, will be appealed by the EU, and will likely drag out for many months, as had the €746m GDPR that was levied against Amazon (AMZN)  in July 2021 (still unpaid), although the GDPR has certainly been a major step forward toward the protection of privacy, at least in the EU, and a similar edict should, in our opinion, be legislated in the US.  That said, with everyday partisan politics in the US, it is hard enough to keep the country from defaulting on its debt, let alone pass legislation that might cause major companies in the US to think before they mine every scrap of data and sell it to whoever might pay for it.  We can dream, right?

​Universal Display (OLED) indicated that they will be showing, for the first time, a demo of their OVJP (Organic Vapor Jet Printing Systems), which the company has had in development for a number of years.  Based on a series of patents developed by the University of Michigan, University of Southern California, and Princeton University, with whom UDC has IP sharing agreements, particularly with work done by Dr. Mark E. Thompson and Professor Stephen Forrest, early OLED researchers.   OVJP differs from more typical OLED material deposition methods which use fine metal masks, essentially screens, to pattern OLED materials. OVJP heats the materials into a vapor, mixes the material with a carrier gas, and sends it to a print head that precisely places the material. 
As opposed to ink-jet printing, which mixes the OLED materials with a solvent, OVJP uses existing vaporization technology and direct printing, but does not liquify the OLED material, nor mix it with a solvent.  OLED materials that are miscible in solvents have different characteristics than immiscible OLED materials, which means formulations must be tailored to IJP systems, while OVJP can use the same OLED material sources as current mask-based deposition systems.  While the idea of OVJP has been around for some time, it has never been commercially developed until Universal Display took up the cause officially in 2014 when it created OVJP Corp. to design and build a commercial OVJP system.
OVJP is run by Jeff Hawthorne, former CEO of Photon Dynamics, a producer of automated inspection equipment for the display space that was acquired by Orbotech for $290, who was later purchased by KLA-Tencor (KLAC).   Company is still in the early stages of OVJP development, with the development of modules that will eventually become an alpha system, but the company will show its first 200mm x 500mm OVJP demo on a glass substrate, along with a 7-layer mono-color phosphorescent OLED device that was made-up on an OVJP R&D system, which will point to ‘proof-of-concept’ for the process.  The next step in OVJP development will be assembling an alpha system of size, likely about 6x the surface area of the current demo, and eventually scaling the system to commercial production sizes.
As the OVJP concept is one that is based on two existing technologies, it is not a far-fetched concept and essentially takes the best parts of both.  However there is substantial engineering involved and considerable competition from other deposition technologies and alternative materials, which are also in development.  The risk to UDC on OVJP is that other developments obviate the need for OVJP before it is commercially viable, so time is certainly a factor in its development, but the demos being shown by UDC do indicate progress is being made which is certainly encouraging.

We have noted that AI is the topic of the day in tech circles recently, but it has been the focus of the private investment community for a much longer time.  Global AI start-up funding has increased from ~$26.6b in 2019 to $52.1b last year, only 20% less that the value of the entire AI market in 2020.    Along with VC funding excitement, which continues currently, a number of AI companies have gone public, some at unusually high valuations.  We were curious to see how those valuations have been maintained as the AI craze continues and were surprised to see that they have not held up as well as one might have thought given the level of publicity.

In the table below, we look at a number of medium to large AI related IPOs and how the valuations have changed.  This is not a complete list, as there have been a number of smaller AI related company IPOs, and we have excluded all foreign AI company IPOs.  We show the latest quarter sales and operating income, and the latest full year sales and operating income, and a brief description of the company.

​Columns 3 and 4 tell the whole story, and with the two acquired companies the valuation slides from ~$57b to ~$14b, a decline of ~75%, although many of these deals were done near a general tech peak in 2021.  As these are typically relatively young companies, they might one day, be able to grow into those lofty AI valuations, but we expect many AI VCs have realized that they were able to cash out of at least some of their AI investments before the cycle began to turn and have been a bit more conservative about valuations since.

Are you worried that Ai might replace you?  Are you thinking about a career change because of AI?  Did you decide to skip taking that English course at the New School because of AI?  These all  ight seem far-fetched concepts but there are those who have already felt the touch of AI, and its not a kind one.  While robotics were supposed to put a major portion of the population out of work years ago, the workplace evolved over time and robotics has taken a much longer time to find its was into the workplace than originally thought.  That said, AI has far fewer mechanical and physical limitations that robotics, which implies it will entangle itself in human society more quickly.  AI has tremendous power, albeit based on human algorithms, but ene in its present early state, it can do lots of things more quickly and efficiently than humans.
We note that two advertising companies in South Korea have both decided to completely stop outsourcing creative design, proposal writing, copywriting, and all other freelance work, in order to use AI related tools that can accomplish those same tasks less expensively and more quickly, under the guidance of its own employees.  A large noodle company (unknown which) has also decided to give up its outsourced projects, at least at the draft level, as it found an Ai system was able to give it more options for illustrations in less time by using keywords than outsourcers.  The company estimates they can save ~2% by just using Ai at this relatively basic level.
At this stage, unless the Ai is highly trained to a specific dataset, results are good, but lack the creativity of a human, at least a human who has some level of creativity, but there are many ‘draft’ jobs in media that do not require creativity, and many of those have already been given over to bots that take digital news feeds and work them into short headlines and copy that are part of global news feeds.  They are far from creative and anyone familiar with the process can usually spot them quickly, but they will get better over time as will AI illustrators and similar media-oriented content production, and they can work 24/7 without coffee or conversations in the break room.  It would seem that the run-of-the-mill copywriter, illustrator, or video clip editor is going to get a run for their money by AI in the near-term, as editors can request an Ai to give 2 or 3 versions of everything it writes, allowing the editor to choose the best, until one day the division manager replaces the editor with an Ai that chooses which is best, and so on up the chain.  It seems wonderful to think that someday we can all relax and sit in our backyards letting a massive AI infrastructure do all the work for us that is until we lose that creative spark that created AI in the first place.

​Whether you agree that it has been warranted or not, there is no doubt that Chinese semiconductor manufacturers have been severely sanctioned by the US government, increasing the challenge facing the Chinese government’s desire to be a world leader in the semiconductor space.  While China needs US tools and technology to maintain a competitive stature in the world of advanced silicon products, it has done little but lodge complaints, at least up until last week, when China’s National Security Review Office conducted a ‘security’ review of products produced by Micron (MU) that are sold in China.
According to the review, the Micron products did not pass the review and ‘pose a major security risk to the country’s critical information infrastructure supply chain’.  According to Chinese Network Security law, operators of critical information infrastructure in China should stop purchasing Micron products.    The office added “China firmly promotes high-level opening up to the outside world.  As long as it abides by Chinese law and regulations, companies from all countries and various platforms are welcome to enter the Chinese market.”
The US DOC responded with, “We firmly oppose restrictive measures that have no factual basis., while Micron was said to have received notice from the China Cyberspace Administration concerning the review conclusion and would evaluate the conclusion and follow-up options but fell short of acknowledging that it would file a complaint, indicating “We look forward to continuing our discussions with Chinese authorities.”  As Micron is the world’s 3rd largest memory supplier, behind Samsung Electronics (005930.KS) and SK Hynix (000660.KS), with a ~25% share, and generated ~$3.3b in sales to Chinese companies last year, the impact will be felt in the US, while South Korean companies will likely pick up much of the slack.
China is certainly in a less advantageous position than the US when it comes to semiconductors, as it remains years behind the US and others in terms of semiconductor technology and is further limited as to advanced lithography equipment sanctions, and there is little chance that the Chinese government will hit its goal of 70% of Chinese demand being satisfied by Chinese producers.  That said, China’s internal production of its semiconductor memory demand has increased from ~8% in 2020 to 10.7% in 2021 and 16.2% last year., and even with the US ban on DUV tools, China can still buy tools for memory nodes 11nm or above.  Last year 28% of memory sales in China were 14nm or below, which implies that ~75% were above 11nm, which justifies considerable spending by the state and local governments to fill demand.  It seems the Chinese government is feeling sure enough that it can fill the Micron gap with other (non-US) vendors, and eventually fill it itself that it is willing to give the US a taste of its own medicine, a ban with little technology substance and lots of enmity and politics.

Earlier this week we noted that while OpenAI’s (pvt) ChatGPT is a general-purpose Natural Language Model (NLM), adding front-end code that references specific datasets can create systems that can be tailored to more specialized applications, opening up NLMs to the real world of data analysis.  These applications can be specific to an industry, a product type, or customer information, but they all share one common issue and that is they must be reliable.

ChatGPT and Google’s (GOOG) Bard are subject to inconsistencies in their responses ranging from incorrect information, incorrect conclusions, and referencing information that does not exist.  Data scientists are not sure why large NLMs have this issue, but we have faced such situations even in our limited use of NLMs.  ChatGPT, given that its training ended in late 2021 is careful to point out that its data is historically limited to the date on which its training ended and might not have more current information, despite its insistence that it has been updated a number of times since then. Google’s Bard however indicates that it is constantly being updated and states that ‘My training is never finished’, leading one to believe that its data is up to date, yet we find inconsistencies in Bard’s responses that indicate a lack of self-checking that is essential for use in most business circumstances.

Here was our simple test:

 Q: How many VR headsets were sold in 2022?

A: According to the International Data Corporation (IDC), global shipments of AR/VR headsets declined 20.9% year over year to 8.8 million units in 2022.

Q: Can you break that down by quarter?

A: Sure. Here is a breakdown of global shipments of AR/VR headsets by quarter in 2022:
Quarter                           Shipments (in Millions)
1Q                                   1.6
2Q                                   2.3
3Q                                   2.7
4Q                                   2.2
 
Q: Can you break that down by brand?
 
A: Sure. Here is a breakdown of global AR/VR headsets by brand in 2022.
 
Brand                              Shipments (in Millions)
Meta                               4.6
HTC (2498.TT(                 2.1
Sony                                1.1
Others                             1.0
 
Q: Can you break down the full year by quarter for each brand?
 
A: Sure.  Here is a breakdown of global shipments of AR/VR headsets by brand and quarter in 2022.
 
Brand                              1Q         2Q         3Q         4Q
Meta                               1.4        1.8        1.7        1.7
HTC                                 0.6        1.0        0.5        0.1
Sony                                0.3        0.3        0.3        0.2
Others                             0.3        0.2        0.2        0.1
 
If we add the quarters in last set, the full year comes to 10.7m units and each brand’s yearly total is different from the full-year totals given in the previous questions.  We asked Bard the following question to clarify:
 
Q: Why do the quarterly VR brand numbers differ from the yearly brand numbers?
 
A: There are a few reasons why the quarterly VR brand numbers differ from the yearly brand numbers.
One reason is that the VR market is still in its early stages of development, and there is a lot of volatility in the market. This means that the quarterly numbers can be affected by a number of factors, such as the release of new headsets, changes in the market, and economic conditions.

Another reason is that the quarterly numbers are based on shipments, while the yearly numbers are based on sales. This means that the quarterly numbers can be affected by factors such as inventory levels and the timing of shipments.

Finally, the quarterly numbers are often based on estimates, while the yearly numbers are based on actual sales data. This means that the quarterly numbers can be more volatile than the yearly numbers.
Despite these factors, the quarterly numbers can still be a useful tool for tracking the VR market. By tracking the quarterly numbers, you can get a better understanding of the overall trends in the market and the performance of individual brands.
​
Here are some additional factors that can affect the quarterly VR brand numbers:

• Seasonality: The VR market is cyclical, with sales typically peaking in the holiday season. This can lead to higher quarterly numbers in the fourth quarter of the year.
• Marketing campaigns: VR brands often launch marketing campaigns in the run-up to major events, such as the Consumer Electronics Show (CES). This can lead to higher quarterly numbers in the quarter leading up to the event.
• New product releases: VR brands often release new products in the first quarter of the year. This can lead to higher quarterly numbers in the first quarter of the year.
• Economic conditions: The VR market is sensitive to economic conditions. When the economy is doing well, people are more likely to spend money on discretionary items like VR headsets. This can lead to higher quarterly numbers in economic boom times.

Overall, the quarterly VR brand numbers can be a useful tool for tracking the VR market. However, it is important to keep in mind the factors that can affect the numbers and to use them in conjunction with other data sources to get a complete picture of the market.

If we had not asked the clarification question or fact checked the answers, we could have based a decision on information that was not correct. The system’s initial reasoning for the differences in its stated results (market volatility) was suspect as it had no bearing on why simple math produced different results from facts that the system presented originally.  The 2nd explanation (some were shipments, some were sales) was inconsistent in that it was not specified in the initial answer, an absolute necessity, in both the actual questions and the system’s specific answers, which were always said to be shipments by the system for VR   Only when questioned did the system indicate that some of the numbers were referencing sales rather than shipments. 

Additionally the system seems to use an ‘answering’ technique often employed by politicians when asked questions they do not want to answer or cannot answer, that of adding a hefty supply of information that does not directly relate to the answer, hoping that the additional verbiage will mask the fact that the answer was not given  The scary part of that theory is that such a technique is a very human trait that had to be programmed into the system or ‘learned’ by the system after seeing the practice used in its training data.  Maybe we are anthropomorphizing Bard a bit too much, but picking up bad human characteristics is not something we hope for or expect from AI.

Regardless, no matter what the circumstance, a NLM is only as good as its training and inference algorithms and its training data, which leads us back to the point that no matter how large the data sets or how specific they are to the required answers, the accuracy, precision, and recall parameters (also known as the ‘F1’) is the key to understanding the reliability of the system.  When humans make decisions based on information presented, they also have an F1, with successful humans having a high F1, while those with low F1s are less successful when it comes to decision making.  As each human programmer has an F1, the accumulated F1’s of the programmed algorithms would be an integral part of the system’s F1, so we can’t blame the system for coming up with incorrect answers.  Perhaps a year backpacking across Europe might help Bard to gain some needed perspective and get down to finding its own personality and consistency.