​Apple (AAPL) devices are among the most fertile fodder for speculation and rumors, with everyone from ‘leakers’ to analysts citing “Everything You Need to Know About the (insert Apple Product Here)”, with much of the information meeting the analytical standard of ‘…even a stopped clock is right twice a day…’, but the media needs content and to many rumors are content, correct or not, so the mill continues to grind out speculation on a daily basis.  As we have noted previously, Apple did not mention anything about the potential release of an AR/VR/XR product at the last developers conference, to the disappointment of many who were looking for some indication that such a device was imminent or at least in a later stage of development, however that has done little to stymie the flow of speculation about such a potential Apple release, in fact it seems to have increased the tide of conjecture.
Most recently there has been some specific reference to the Apple AR/VR/XR supply chain, which, according to some, is gearing up in anticipation of a product (aka ‘Project N301’) release as early as the end of this year or early 2023, with a 2nd product (aka ‘Product N421’) slated for 2024.  The first product release is said to be a VR headset with additional AR capabilities, while the 2nd is either an updated version of the N301 or an AR only device.  While we know Apple has a number of display research projects in the works, Micro-OLED seems to be getting the most attention as a VR display for such a project, and where LCoS (Liquid Crystal on Silicon) was the early AR display favorite, that technology has been mostly replaced with Micro-OLED, given the infrastructure that has been developed around small panel RGB displays and to a lesser extent OLED lighting. 
While the substrate for smartphones is a flexible plastic or glass, Micro-OLED displays are constructed on Silicon substrates and can therefore utilize semiconductor tools for many of the associated processes, and through systems such as multiple stacks and direct patterning, high resolution and high brightness OLED displays can be produced.  With typical small (1” to 2”) OLED displays being very low resolution (450x450) and relatively low brightness (350 to 450 nits), such displays are not suitable for AR or VR applications, which require resolution of 1920x1080 at a minimum and brightness levels of multiple thousands of nits, and as earlier AR/VR displays used WOLED display technology that required a color filter, such levels were not achievable.  As Micro-OLED uses patterned RGB emitters, it requires no color filter, which can vastly improve brightness.
But the problem with Micro-OLED technology is it is relatively new and each company developing the technology has their own way of patterning and driving the OLED materials, which means an ‘un-common’ infrastructure and supply chain.  While an AR or VR device by Apple will go a long way toward developing a supply chain around Apple’s BOM and chosen technology, we expect that technology to change very rapidly, making the development of a mass market infrastructure hard to develop., especially if Apple’s technology differs from that of Meta (FB), who is the current share leader in the VR headset space. There are a number of companies that produce Micro-LED displays, some of which are focused on the display side while others are dedicated to projection modules but each has its way of differentiating itself from competitors.  
Large display panel producers like China’s BOE (200725.CH) or LG Display (LPL) or well-known image sensor manufacturers like Sony (SNE), and more specialized OLED producers like eMagin (EMAN) and Kopin (KOPN) are all working toward the goal of finding an inexpensive way of producing Micro-OLED displays to feed the XR space, particularly AR development, but Apple itself is deeply involved in such development, having applied for or been granted over 53 patents related to AR and 51 related to VR just in the US, going back as far as September 1999 and as recently as last month.  While many of these patents have to do with the mechanics of display image processing or optical systems that bring images to the user’s eyes, Apple is deeply involved and likely has the goal of producing some device that they can sell and provide content for. 
Whether the device appears this year, next year, or in 2024, Apple’s ‘legitimization’ of AR and VR technology to the average consumer will be a big step forward for the industry, which is currently quite fragmented, but as we have said in the past, the key to really expanding AR usage is the creation of applications that will drive consumers toward the technology, not because Apple says its good and they can show off a well-designed AR device, but because the applications become ingrained in our culture.  Selling product will be the focus for retail applications and location services will certainly be enhanced, but an application such as language translation would be a killer application in our view.  Imagine walking down the street in a foreign country and just by looking at a person, you could understand what they are saying in any language.  Of course the quality of the translation systems are a key, but just the ability to get an understand of local culture by listening to everyday conversations would be a benefit to all, and that’s only one application, so if Apple really wants to sell AR products they should be spending mucho dollars on application development, as without effective consumer applications, the best you get a a good-looking headset that sits in a drawer.

​On July 8 bad weather in Hiroshima, Japan caused a power outage at Micron’s (MU)  Fab 15 DRAM plant that lasted between 5 and 10 minutes, and while backup generators were triggered, a portion of the production tools had to be restarted and recalibrated.  Fab 15 accounts for ~30% of Micron’s monthly wafer starts and ~7% of the global total and is focused on smartphone DRAM and R&D.  This happened roughly a week after Renasas (6723.JP) was forced to shut down its Kawasaki fab for a week after a lightning strike that hit a power line outside of the plant.
Micron cautioned that the shutdown will have some effect on quarterly results but was still assessing the damage to wafers that were on the line during the outage, but the impact on the overall DRAM market is expected to be minimal as inventory levels are high and the industry is facing continued oversupply.  Given that Micron will be working at low production levels and will likely have to scrap partially produced wafers, the resulting production losses will likely work toward bringing down inventory levels, but not enough to make an impact on pricing, other than in the spot market, with expectations for a ~10% price reduction for DRAM in 3Q remaining intact..

While Samsung Electronics (005930.KS), might not have been the first to release a foldable device, they certainly have been the category’s biggest proponent.  With the company expected to release its 3rd generation foldable series next month, there is little about the Z4 line that is not already known, other than price, although speculation is that both models will remain at the same price as the previous iteration.  That said, what about the Galaxy Fold 5 and Flip 5, which will not be released for over a year?  Yes, there is already the beginnings of conjecture about that next generation of foldable smartphones, even before the previous version is out.
The Galaxy Fold 5 and Flip 5 are expected to be based on Qualcomm’s (QCOM) Snapdragon 8550 chipset (aka Snapdragon 8 Gen 2), a 4um node device produced by Taiwan Semi (TSM) to be released later this year as an upgrade to the Snapdragon 8+ Gen 1, which was recently released.  This has also brought on speculation that Samsung will no longer use its own Exynos processor in the foldables (and potentially other mobile devices) as the Snapdragon specs have bypassed those of the Exynos.  But there is one aspect of the Flip/Fold 5 that is unusual, and that is Samsung’s unit volume targets, which are expected to be quite low and are seemingly reflecting a dour picture of the smartphone or CE environment in 2023.
As shown in the table below, we believe Samsung shipped 7.11m foldable units in 2021.  Samsung originally set a target of 10m units for this year but increased that to 15m units earlier this year.  Thus far we have not heard that the target for this year has changed, and, if the release date of August is correct, production of those units has been underway since late May or early June, especially considering the difficulties associated with shipping that CE companies are currently facing.  As production for the Galaxy Fold/Flip will continue through the end of the year, there is stil time for Samsung to reduce the number of units once they see whether the macro environment is affecting foldable sales. 
That said, the trade press in Korea is stating that Samsung has already set its unit volume targets for the Galaxy Fold/Flip Series 5, tobe released in August 2023, but at a considerably lower target (10m units) than one might expect given Samsung’s hopes for growth in the foldable smartphone category. This could be a reflection of Samsung’s 2023 macro outlook or more likely in our view, a very conservative estimate that will likely be raised, a more positive timeline than one where targets have to be reduced.  We note also that the ratio between the two models in both 2021 and 2022 has been 3:7, while in 2023, it is 1:4, according to the same sources.  This implies that the Galaxy Flip 5 will see an increase in unit proportion against the more expensive Galaxy Z Fold 5. 
Again, without confirmation from Samsung, we expect this would reflect an anticipated decline in price for the Galaxy Z Flip, which would increase its customer base, while the Galaxy Z Fold remains near or at its current levels to preserve margins across the entire foldable smartphone line.  This also does not assume that Samsung will release any new foldable smartphone models, which we believe is certainly a possibility in 2023, so the speculation of over the Z Flip and Fold targets could be excluding other models that would bring the targets up to what would be more ‘growthy’ levels, especially as Samsung has stated that it expects to expand the foldable line over time.

​Vivo (pvt) is not a smartphone brand well-known in the US, however the company held a 7.9% share of the global smartphone market at the end of last year, and a ~16% of the Chinese market at the same point.  Owned by China’s  BBK Electronics (pvt), along with its sister brands Oppo (pvt), OnePlus (pvt), RealMe (pvt), and iQOO (pvt), Vivo is among the most popular smartphone brands in India, with a share that is on par with Samsung.  That said, there has been considerable enmity between the Indian and Chinese governments over border issues for the last few years and occasional flare-ups of anti-Chinese sentiment have been focused on Chinese smartphone brands at times.  Some of that antagonism seems to have made it way to the Indian Law Enforcement Agency who recently froze Vivo’s bank accounts, citing the investigation of 48 Vivo locations in India under provisions of the Indian Money Laundering Act.
The Bureau has accused the company (and others) of remitting 62.48t rupees (~$7.85b US) to China and other locations to avoid paying taxes, which represents about half of the company’s revenue.  The bureau froze 119 Vivo accounts totaling $57.4m, along with fixed deposits, 2Kg of gold, and cash.  Vivo has asked the New Delhi High Court to unblock a number of the accounts so it can pay salaries and dues, who then gave the enforcement agency until tomorrow to make a decision as to whether to unlock the few necessary accounts.  The enforcement agency has indicated that the capital was shifted back to China to make the Indian operation look like it was producing a loss and therefore avoid taxes and alleged that some Vivo executives had tried to leave the country when they discovered the investigation and had used forged identification and address documents when they formed the Indian subsidiaries of the company, strangely using the addresses of what turned out to be a government building and the home of a local bureaucrat instead of the real corporate address.  The investigation continues….
 

While Prime Day is not considered in the same league as Alibaba’s (BABA) ‘Singles Day’, it does tend to be a significant selling day (actually 2 days) and generates roughly the same in sales for Amazon (AMZN) as does ‘Cyber Monday’ for a large group of retailers and CE companies.  There are many tricks that retailers use when reporting numbers for such shopping holidays, so we take most of the numbers with a grain of salt, but no matter how much the numbers vary from source to source, they are significant, although dwarfed by Singles Day, which is based on live streaming (sort of an internet QVC) and generates 6 to 8 times more revenue, depending on whose data you use.
Since we have been tracking Samsung’s (005930.KS) Mini-LED/QD TVs since they appeared last year, we took another look at the 2021 and 2022 Samsung ‘Neo’ and ‘QD’ TV lines to see if Prime Day had any significant effect on prices.  That said, the TV industry is currently in an odd position in that TV panel prices have fallen to near pre-pandemic levels, which should translate into much lower TV prices than seen last year, but at the same time other components remain at peak or near peak levels and the cost of transportation has become another cost that has had significant negative impact on the TV supply chain.  Coupled with inflation, and 2020/2021 consumer pull-in purchases, the prospects for a robust pre-holiday season are less than ideal.
As Samsung has both 2021 and 2022 Mini-LED/QD models, we break down the data into yearly segments with the 2021 line seeing a 1% increase in price since the last check (18 days ago), while the 2022 line saw a drop of 4.7% during the same period.  One unusual factor that appeared in both 2021 and 2022 lines was the fact that for the first time the two 98” Mini-LED/QD TVs in both lines, which had been priced at $15,000 each, saw the first discount this week since they have been offered.  The new price of $13,000 was a 13.3% discount, without which the 2021 line would have seen a 1.7% overall price increase (rather than a 1% increase) and the 2022 line would have seen a 4.0% decline rather than a 4.7% decline.
Notwithstanding the 98” TV discounts, the 2021 line had 13 of 36 models (36.1%) at their lowest ever price and 1 (2.7%) at its highest ever price, with 7 models up and 6 down since the last check.  The 2022 line had 30 (85.7%) at their lowest prices and 4 (11.4%) at their highest since release, with 3 up and 21 down.  We took the comparison a bit further in the table below which shows the following:

• The Model, Resolution (8K/4K), Features, and display size
• The Initial Price of the 2021 and 2022 sets and the difference if any
• The Current Price of the 2021 and 2022 sets and the difference between that price gap
• The discount between current price and initial price for both years
• A timeline check where we match the current (2022) model prices with the 2021 model prices at the same point in time, which in this case is 90 days from the 2022 release date.  This indicates whether the discounting is occurring more or less aggressively than last year.

 
Here are our conclusions:
 

• The initial offering price of the 2022 Mini-LED/.QD and QD only line, excluding those models that were not offered in 2022, was 4.3% lower than the equivalent sets offered in 2021. 
• Currently the 2022 models are 19.1% higher in price than the 2021 models in aggregate.  The 2021 models, which have been out for 418 days are priced 29.3% below their initial price while the 2022 models, which have been out for 90 days, are priced 12.6% below their initial price. 
• If we look at the 2021 line 90 days  after their release, there are 14  (45.2%) 2022 models that are discounted more heavily than  at the same point last year, 15 (48.4%) that are  discounted less than last year, and 2 (6.5%) that are discounted at the same level as last year’s models.  This leads us to conclude that other than the initially lower price of the 2022 models, Samsung has not been any more aggressive in discounting the 2022 models this year than last, despite the competition in the Mini-LED/QD TV space.

We believe that while Mini-LED arrays have come down in price over the last 400+ days, along with LCD panel prices, the price of other components has increased, along with transportation costs, but given that an LCD display open cell is ~25% of the total manufacturing cost of a TV, the overall   cost of the ‘other’ manufacturing components and processes remains an offset to the panel price reductions.  That said, there are certainly margin considerations here, and with Mini-LED/QD TVs being considered ‘premium’ TVs, they will likely carry higher margins for Samsung than more generic LCD TVs, which means Samsung needs them to preserve overall TV margin. 
While TCL (000100.CH) is a serious price competitor in the Mini-LED/QD TV space, LG Display’s (LPL) Mini-LED/QD line is less so as it is more closely priced to the Samsung line, with much depending on the particular market and whether TCL is a presence.  In the US we believe Samsung  offers 3x the number of overall TV models than TCL and ~30% more than LG, and while this does not equate to unit volume or sales directly, it does give some indication as to the  choice between price reductions or marketing dollars for Samsung.  With Samsung holding a 20%+ share of the TV market and TCL under 10%, the decision to hold price seems one that Samsung has made years ago, especially on its premium products, so we expect, while there will be additional discounting on the 2022 line going forward, just as there was on the 2021 line last year, Prime Day is not the inflection point one might expect for the Mini-LED/D TV world.

​AGC (5201.JP), aka Asahi Glass, a competitor to Corning (GLW) and NEG (5214.JP), has raised their 1H sales and operating profit guidance.  Sales, which were original forecasted to be ¥870b yen, have moved up by 11.5% to ¥970b and raised its 1H operating profit forecast by 21%.  The new sales forecast is 7.2% above last year’s 1H, while the new Op Profit forecast is up 20.8% over the same period last year.  The original forecast was made by the company on 2/8/22.  While we were hoping that the increase might have been associated with the company’s display glass business, Asahi indicated that the company’s caustic soda (Sodium Hydroxide – NaOH – aka lye) business was positively affected by price increases in the chemical, which is used in paper processing, soap manufacturing, and in the refining of aluminum, along with a number of food preparation processes.   The company also indicated that its PVC (polyvinyl chloride) business, among the top 3 synthetic polymer plastics produced globally, saw higher than expected prices in Southeast Asia during the 1st half.  The company was also able to pass on the higher cost of raw materials and fuel prices to customers in its architectural glass business in Europe.  No mention of the display glass business.  We will likely have to wait until the call next month.

The smartphone market has been relatively weak this year, with optimistic shipment targets made late last year being trimmed so far this year.  There has been little hope that sales will pick up appreciably, outside of seasonal norms in 3Q and 4Q, other than Apple’s (AAPL) iPhone 14 family release in mid- September, and that could be delayed a bit if there are problems with logistics.  Typically (5 year average) smartphone shipments increase ~1.1% in 2Q and 13.5% in 3Q, and while much of the hard shipment data for 2Q has yet to be published, early estimates are that shipments are expected to decline ~3.0%.  3Q however is more of a crapshoot (non-technical term for statistical anomaly)) as the typical 3Q q/q gain is 13.5%, which, given the macro environment, might be a bit difficult for the industry to meet, especially given the higher than normal inventories plaguing a number of CE product categories.
In the past OLED based smartphones have defied broader growth patterns for the broad smartphone market as share grew, but with OLED smartphone share reaching ~43% last year and continued share growth this year, OLED smartphones are mainstream and face macro challenges similar to those for LCD smartphone displays.  This is made more understandable considering that the competitive nature of the OLED display space continues to increase with Chinese small panel OLED display producers pushing to gain share to prove their worth in the overall display market, as they have done in the large panel LCD space and Chinese rhetoric about how they are challenging the dominance of South Korean small panel OLED producers Samsung Display (pvt) and LG Display (LPL) with a bent toward the goal of ‘world dominance’.
As we have previously noted, South Korean small panel OLED producers have a number of advantages over Chinese small panel OLED producers, primarily the production experience that comes from 15 years of small panel OLED fab operation and a broad supporting infrastructure, while Chinese players began what would barely be called commercial production 9 years ago.  That said, Chinese small panel OLED producers do have one large advantage, and that is capital, which despite questionable operational profitability (without subsidies), seems to be readily available to OLED producers for expansion and to make up operational shortfalls.  The expansion capital has allowed Chinese small panel OLED producers to add considerable capacity over the years as seen in Figure 2, and without question we expect Chinese small panel OLED producers to eclipse South Korean capacity some time over the next 2 – 3 years.
That said, while the numbers would indicate that Chinese small panel OLED producers can take over the lead position in production capacity from South Korean rivals, there are missing factors in Figure 2, and those are utilization and yield.  Given Samsung Display and LG Display’s long-term small panel OLED production experience, they maintain high yield ratios, and while they vary according to how long a product has been in production and the complexity of the display, we believe South Korean small panel producers maintain higher yields than their Chinese rivals and are therefore able to produce more sophisticated small panel OLED displays.  While such sophisticated displays carry a higher profit margin, the market for same is smaller than that of more generic small panel OLED displays, which theoretically should lead Chinese small panel OLED producers to higher utilization rates and higher unit counts, however this is not the case based on our data, and even with China’s largest OLED producer BOE (200725.CH) becoming part of the Apple iPhone OLED display supply chain, utilization rates at Chinese small panel OLED producers remain low, and given the weakness in the overall smartphone market, we expect relatively low utilization rates to continue into 3Q for most Chinese small panel OLED producers.
We expect composite small panel OLED demand to be up 8.7% q/q, but down 1.8% y/y, with only 4 of 8 smartphone brands (including ‘other’) seeing an increase q/q, with Apple the greatest increase based on the iPhone 14 estimated release dat.   We do expect declines in small panel OLED displays from Huawei (pvt), Oppo (pvt), and Vivo (pvt) in 3Q, which would affect production at Visionox (002387.CH) and Tianma (000050.CH), while any Chinese branded smartphone order reductions at BOE will likely be offset by production for the iPhone 14, for which BOE is expected to produce 5m units.  To put that in perspective BOE has been producing ~15m small panel OLED units on average, for the last 6 quarters, with SDC and LGD expected to produce 60m and 25m respectively, based on a 90m order from Apple, which seems to be the latest estimate.
While we are certainly not denigrating the competitive threat to South Korean small panel OLED suppliers from Chinese suppliers, we point to the fact that even with BOE’s participation in the iPhone 14, Chinese small panel OLED producers have maintained a relatively stable supply share of ~23.6% (6 quarter average including 2Q/3Q forecasts) while South Korean producer share has averaged 76.4% over the same period.  Given our expectations for 23.2% and 76.8% respectively for 3Q, we see little change over the average.  As we noted above, we do expect China’s small panel OLED share to increase as their capacity and experience grows, but we are far more willing to accept that change when factoring in utilization, yield, and profitability and given how rapidly the technology seems to be changing, we expect it will take more than subsidies and capacity for Chinese small panel OLED producers to overtake South Korean suppliers in the near to mid-term.  After that, it’s anybody’s game…
 

​The headlines in a South Korean tech paper indicated that a research firm had discovered that Samsung Electronics (005930.KS) was selling its 65” QD/OLED TV (S95B) at a price $200 below that of its main competitor LG Electronics (066590.KS) 65” W-OLED TV (Model G2), which would imply Samsung is willing to keep this new technology priced as a direct alternative to LG’s OLED TV products, regardless of how LG prices its comparative products.  With the original 65” QD/OLED price of $3,000, Samsung has already reduced the price, which remains at $2,600 on their website and within a few dollars of that more major retailers.  That said, we did find one on-line vendor that was offering the unit for an unusually low price of $2,049 out of Hollywood, Florida, although neither sales nor support lines seem to be working.
While we do expect that Samsung will try to remain competitive with LG’s OLED TV offerings with LG’s G2 series priced at $1,999 (55”) and $2,599 (65”), we were unable to find the offer mentioned in the trade press, although we expect there will be a few ‘door buster’ type sales over the next few months that might get the price a bit below the LG range, but we expect Samsung has built-in some substantial controls as to how the QD/OLED TVs are priced in order not to create unfair competition between vendors.  We believe it is frowned upon for brands to officially lock retailers into a price but we know of a number of instances where large brands have castigated those who beak the ‘suggested’ price points.  We do expect that Samsung will have some room to lower QD/OLED prices as it improves yield, and we would expect the company is still experimenting with how it is going to present the new technology from a pricing standpoint, so there might be a point where Samsung is willing to sell for little or no margin to make sure it sells out the targeted number of units before year end.  We are still watching…

​According to Korea’s ETNews, Apple has confirmed that the rumored OLED iPad will be released in 2024.  Apple and suppliers have been producing prototype of the device to make sure that the OLED displays can scale to while still meeting Apple’s specifications as the OLED iPad, should it be similar in size to the current model, would have a screen with almost 3x the size of an iPhone.  With sales in the 50m to 60m range, a mistake on the company’s first OLED base iPad would be quite expensive and damaging to the company’s reputation and the extra time the company is taking to make sure things are right is similar to how the company worked through the iPad OLED conversion years ago, which was slower than most major smartphone brands and started with one model, eventually expanding to the entire iPhone line.
There is still considerable speculation as to who will be supplying the displays for the OLED iPad with Samsung Display and LG Display in primary contention, with both indicating potential capacity expansion plans geared toward producing larger OLED panels on Gen 8 substrates rather than the more typical Gen 6, which would lead to more efficient processing.  China’s BOE is also angling to make its way further into Apple’s display supply chain, although the company is still behind SDC and LGD in terms of producing acceptable yields on LTPO displays, which have been the backplane technology of choice for Apple’s high-end iPhone models.  With the iPad typically being released in September, BOE and potentially others, will have a bit over a year to supply Apple with additional prototypes and gear up production, but it will be difficult for new fabs to be built and tested in that time, so we expect the first round of OLED iPads will still be produced o Gen 6 lines with the possibility that there will be more than one supplier, even if there is only one OLED iPad model in the 2024 year.
 
 

​There is a war going on that rarely makes it into the press as the battlefield is not on the ground, in the air, or in space, but inside the guts of massive processing nodes that are used to understand the nuance of language.  The algorithms that direct these processor are based on a subset of Ai that deals directly with syntax, expression, and a host of other language variables that make understanding other languages a challenge for humans and a monumental task for digital entities.  The opponents here are companies rather than political entities, and well-known ones at that, with Microsoft (MSFT), Google (GOOG) and Meta (FB) all pitted against each other with the focus of being the dominant force in the digital translation market.
There are no casualties or battle lines in this war, as it is fought with processing metrics and advertising, and that makes it hard to know who is winning, but the participants seem to have taken on a particular metric to make the public aware of who might be in the lead.  That metric is the number of languages a system can translate, which seems to be used a s a gauge toward whose system is the most advanced.  While the number of languages that a system can translate is certainly important, especially to those languages that might be considered secondary, but by far the most important metric for translation services is accuracy. 
Recently Meta indicated that its NLLB-200 Ai model has increased the number of languages it can translate to 200, which it accomplished in two years, while Google Translate is only able to work with 133 languages and Microsoft’s system translates only 111 languages, although that includes two Klingon dialects, and is clearly staking a claim as the world’s most advanced translation tool.  While the number of languages a system is able to detect is certainly easily understood by the general public, the quality of the translation, which is based on the algorithm and the sample base, is far more important and there are two ways in which that can be evaluated, by humans or by machines.  Using humans to evaluate translation quality throws subjectivity into the mix, while automated machine evaluation does not, but a machine evaluates a translation by averaging words and sentences against a human evaluation, with the idea that the closer the machine score is to a human score, the better the translation is.
With all of this being beyond the scope or desire of the general population, translation giants will continue to use the simplest metrics to give credence to their systems, but will have little correlation to real world results, as the ability of the AI to understand nuance and what to do with that understanding is really the key.  All three companies mentioned here have access to vast stores of speech, which certainly goes toward the ability of an AI to learn, but the algorithm is the key and that is something that not only grows with more resources but must change as humans better learn how to convert those subjective views into language that a machine can understand.  So the question is, does the AI need more language to read or does it need more human evaluations in order to take the subjectivity out of their evaluation?  The only way to know is…
E ninau i ka Mīkini - Hawaiian
Spurðu vélina - Icelandic
Faighnich dhan Inneal – Scottish Gaelic
quaeritur machina – Latin
Ask the Machine - English