How long before Google kills British English? (And does it matter?)

I’m (slowly) learning an excellent Chinese Karaoke song. I’ve so far got the chorus and most of the first verse. My process is: I print out the pinyin and the Chinese characters and gradually try to recognise the characters. I was recently looking at 裤子 – here is what Google Translate has to say about these characters:

The translation is “Pants”, or the American English version of the word. If you squint carefully, you can see there is another option for the British English version: “Trousers”, but how many people are going to do that?

It reminds me of a big debate I had with my kids a few years back about what these are called:

They insisted “Ladybug”. Took a while to get them to accept “Ladybird”.

I wonder what Google Translate thinks. Sure enough, if I use Google Translate to translate “Ladybird” to Chinese and then back to English I get “Ladybug”:

Again, with a tangential nod to “Ladybird” but, frankly, the damage has already been done.

Sharing this because it’s an interesting story about changing language: if the tools the people use to translate into English only consider American English then how long before British English disappears? And does it matter?

Let’s try one more for fun: Pavement. Hmmm, what do we have here:

It’s using the American meaning of the word to translate to the “road surface”. Squinters will notice that you can find a “sidewalk”/”pavement” option.

Translate 路面 back into English with Google Translate and you get ‘pavement’ (as in road surface):

“Sidewalk” gets translated to 人行道, which when translated back into English only gives the option for “Sidewalk”, not for “Pavement”:

When a word can have one meaning in British English and one in American English, Google treats the American version as the default¹.

So this is not just about translating into English. It’s also about translating from English.

Seems that from Google’s point of view, American English is the default option and, only if there isn’t an American English meaning of a word will it grudgingly accept that British English exists².

I mean …. where will this end? Language and spellings are one thing… but will Google eventually try to convince the world that the US date format should be treated as the standard? Doesn’t bear thinking about 😀

I wonder whose side ChatGPT is on in this subtle manipulation of language….

Ah, phew. All is not lost. At least ChatGPT hasn’t been absorbed into the Google Borg of American English. Though…. as it trains on more and more text that people have written with Google Translate…. how long will ChatGPT be able to hold out before it is assimilated? Is resistance futile?

Footnotes

1. I am reliably informed by my Chinese friends that 路面 is a reasonable word to use for the whole wider concept of a road, so you could use this word to describe a trip hazard that a person might fall over or a bump that a car might encounter. But evidently Google is leaning towards the sense of ‘where the vehicles go’ ↩︎
2. I’m also reliably informed by my Chinese friends that state-provided schools in (mainland) China tend to use British English books rather than American English. Oh the irony. ↩︎

Knowing Your Onions: Adventures in cross-language misunderstandings

I recently bought a recipe book in Greek. It was a result of spending months looking for a great Kleftiko recipe – eventually I found one that worked with a few small modifications. Anyway, for anyone who wants the recipe this is the one from Akis Petretzikis – but I switch leg of lamb for shoulder. I wonder if lamb cuts vary from country to country – are they leaner in UK than in Greece?

I made another recipe and got one of the ingredients a bit wrong. Here’s why.

‘κρεμμύδι’ (kremidhi) is “onion”. The ending ‘-ακι’ (aki) means “little”, so ‘κρεμμύδακι’ would be a small onion. ‘φρέσκο’ (fresko) means “fresh”, so ‘φρέσκο κρεμμύδακι’ would mean “little fresh onion”. Except, it turns out, it means “spring onion”. No wonder my meal looked a bit different to what the picture in the recipe book showed. I knew the words individually, but not when they were put together.

It made me wonder what it must be like to be working in an English speaking organization if English isn’t your first language (or even if it is). Even if you understand the individual words, there must be so much scope for misunderstanding when you put words together.

Here is the Google Translate evidence for your delectation:

Notes on Sentience and Large Language Models

A collection of some relevant commentary about the recent story that a Google employee [was] reportedly put on leave after claiming chatbot became sentient, similar to a ‘kid that happened to know physics’.

First: the posting of the interview with the AI (it’s called LaMDA). It’s worth a read if you want to see what is sitting under all the fuss.

Second: The unequivocal counter-argument that, while this is a tremendous advance in language modelling, it is not sentience, it is just a massive statistical model that generates text based on text it has seen before:

Nonsense. Neither LaMDA nor any of its cousins (GPT-3) are remotely intelligent. All they do is match patterns, draw from massive statistical databases of human language. The patterns might be cool, but language these systems utter doesn’t actually mean anything at all. And it sure as hell doesn’t mean that these systems are sentient.

Third: The text that the AI generates depends on the text it has seen. It is very responsive to leading questions:

Just had a look at the 'LaMDA sentience' interview in case there's anything at all interesting there, and no it's literally just that language models will go along with leading questions.
Look at how GPT-3 (a similar model) answers the same prompt, and some similar ones: pic.twitter.com/0P6BeT7SOE

— Rob Miles (✈️ Tokyo) (@robertskmiles) June 12, 2022

Is this language model sentient? Is sentience a matter of “yes you are a sentient” (e.g. a dog) or “no you aren’t (e.g. a rock)”. Or are there matters of degree? Is a bee as sentient as a dolphin? If a machine became sentient would humans even be able to recognise it?

We used to think that the Turing test would be a good benchmark of whether a machine could (appear to) think. We are now well past that point.

It takes me back to my anthropology studies. It turns out it’s quite tricky to define what distinguishes humans from animals. We used to think it was ‘tool use’ until we discovered primates using tools. When we saw something that passed as ‘human’ according to the definition, but was obviously not human, we changed the definition. Seems we are in a similar place with AI.

A more pressing issue than the ‘sentient or not’ debate is ‘what are we going to do about it’. It’s the ethical side of these advances which are both terrific and terrifying. With deepfake imagery and these language models, the possibilities both good and bad are hard to get your head around.

So I leave you with.

Fourth:

I like how much AI researchers are like early 20th century nuclear physicists. “This thing we’re doing is really dangerous… and maybe unethical… but let’s keep doing it… unless… oooh it’s so scary… but also cool… are we the new Gods? We’re full of regret and spirituality…”

— Joe Bernstein (@Bernstein) June 12, 2022

Management and product development lessons from the 1950’s

In 1955, Elihu Katz and Paul Lazarsfeld published “Personal Influence“. This studied how small-group dynamics moderate or influence mass media messaging. For example how people decide who to vote for, which brand of lipstick to use, or which movie to go and watch.

Reading this in 2018 it’s striking to see how much is still valid. I’m not posting this to provide tremendous new insights. Any insights here are over 60 years old. Apparently, human behaviour doesn’t change very much over the generations.

How people choose their leaders

In order to become a leader, one must share prevailing opinions and attitudes. (p52)

They cite a 1952 study on children in a day nursery in which kids with “leadership qualities” were separated from the other children who were then placed into groups of 3-6. These new groups created their own “traditions” (e.g. who sits where, group jargon, division of who plays with what objects). The original leaders were then re-introduced:

In every case, when an old leader attempted to assert authority which went contrary to a newly established “tradition” of the group, the group did not respond. Some of the leaders, as a matter of fact, never returned to power. Others, who were successful, did achieve leadership once more but only after they had completely identified with the new “tradition” and participated in it themselves. (p52)

Or another 1952 study amongst a religious interest group, a political group, a medical fraternity and a medical sorority:

[T]hose who had been chosen as leaders were much more accurate in judging group opinion … But this was so only on the matters which were relevant to the group’s interest – medicine for the medical group, politics for the political group, etc. It seems reasonable to conclude … that leaders of groups like this are chosen, in part at least, because of recognized qualities of ‘sensitivity’ to other members of the group. (p102)

A succinct argument as to why people who want to become leaders need to first spend time listening.

Group participation improves take-up

Here’s are some more 1952 studies that the authors cite:

1. A study in a maternity hospital in which “some mothers were given individual instruction .. and others were formed into groups of six and guided in a discussion which culminated in a [group] ‘decision’ [to follow the instruction.” The participants in the group dicussion adhered “much more closely” to the child-care programme. (pp74-75).
2. A study comparing a lecture approach vs a group discussion on “the nutritional and patriotic justifications for the wartime food campaign to buy and serve ‘unpopular’ cuts of meat. 3% of those involved in the lecture followed the desired course of action, vs 32% of those in the group discussion.

Worth bearing in mind in the next meeting you host, or the next corporate communication you send out.

How small groups construct their reality

So many things in the world are inaccessible to direct empirical observation that individuals must continually rely on each other for making sense out of things. (p55)

Apparently 1952 was a bumper year for social sciences. Here is another 1952 study in which individuals were asked to decide how far and in which direction a point of light was moving. The catch was that the point of light was static. The study found that:

1. When people were shown the light individually, they would make their own judgment of how it was moving. When they were later put into small groups of 2 or 3, “[e]ach of the subjects based his first few estimates on his previously established standard, but confronted, this time, with the dissenting judgments of the others each gave way somewhat until a new, group standard became established.”
2. If a group session came first, the group would achieve a consensus of how the light was moving, and each individual would adopt the group’s consensus as their own position.

The way reality is generated by social groups is something to bear in mind during user research activities.

How the make-up of a group affects quality of communication

You guessed it, it’s another 1952 study that found that:

1. Rank in the group affects how people communicate. Specifically: “[P]-erson-to-person messaged are directed at the more popular group members and thus may be said to move upward in the hierarchy, while communication from one person to several others tends to flow down” (p89).
2. As groups get larger (from 3 to 8) “more and more communication is directed to one member of the group, thus reducing the relative amount of interchange among all members with each other. At the same time the recipient of this increased attention begins to direct more and more of his remarks to the group as a whole, and proportionately less to specific individuals.” (pp89-90)

I’m sure these two findings will ring very true of many meetings you’ve been in. I suspect that the person who becomes the centralising leader in these communications might not even realise the role they are playing. Reading this makes me more keen to try out the kind of silent meetings approach they use at Square.

 

 

Some ramblings on culture

Confused of Calcutta had a great post on cultural differences some weeks back. In it JP references this post: http://www.adinochang.com/archives/chinese-culture-versus-german-culture.html. Follow the blog comments and you’ll find a similar lighthearted comparison of Italy and the EU (by which presumably they mean Northern Europe) http://www.lifeinitaly.com/flash/

Do check these 3 links out – they are fun and can teach you a think or two.

In my case I am half English/half Greek. I was born and grew up in England with the exception of a few years in Greece. Though when I say “half Greek” I should clarify: My mother grew up in Alexandria in Egypt and moved to England in her late teens. So her (and therefore my) view of what it is to be Greek is based on a version of Greece that probably diverged from mainstream Greek culture a century ago. Still – growing up in England in a household that frowned on going to the pub and getting drunk (in some versions of Greece, young people will stay up all night eating ice cream, seriously) made me feel like something of an outsider for a great part of my life.

Which is why I enjoy these cultural comparisons so much, I suppose. Because they help teach me about my own culture(s) as much as about other culture(s). For example I identify with a lot of the “Chinese” cultural elements in Adino’s post, particularly around the significance of the family, but in other respects I am much more “German”. These kinds of jokey stereotypes help crystallise some important cultural points.

So far so rambling, but there is something here about technology and business.

To date, “globalisation” has essentially meant the spread of North American firms and culture. But there is no way that this flow of globalisation can continue one-way. According to the BBC here it is expected that within 2 years there will be more internet users in China than in the USA. So not only does it make sense for westerners to make more efforts to understand cultural differences, but westerners shouldn’t be surprised if what passes for “normal” on the internet in China becomes more dominant globally. Perhaps the next generation of internet sites will look more like TaoBao than eBay.

Buying in Ancient Times

Enjoyed reading Purchasing at the Parthenon by Bob Soames in a recent Supply Management article (subscription required for online access). It describes procurement practices in ancient Greece, 2,500 years ago – the challenges they faced then were very similar to the ones buyers face today: cost overruns, fraud, poor quality supplies etc.

Two interesting solutions of theirs:

The city of Ephesus contracted its works out to private companies. When agreeing a project, the architect had to sign over all his property to the city as security. If the project came in under budget then all well and good. The city would even pay for budget overruns up to 25% of the agreed budget. Anything beyond a 25% budget overrun would be paid for ourt of the architect’s own purse.

Buyer beware: during those times a distinction was made between business-to-consumer and business-to-business deals. In B2C sales, the buyer was allowed to return your goods as not fit-for-purpose. In B2B, however, professionals were expected to manage risks themselves.

Participant Observation and Software Product Development

First of all there was the “them” and “us” of “IT” and “The Business”, the methodology of choice was Waterfall and IT’s job was to implement what the Business said it wanted. Projects were late, expensive and disappointed everybody.

Then came Rapid Application Development, DSDM, UML, Rational Rose etc. On of the big ideas here was to get “IT” and “The Business” into the same room so they could decide together what they wanted and hopefully come up with a better result. Thankfully it was better than the Waterfall approach.

Where next?

The next evolution, I hope, will be be building software through participant observation. Anyone who has done much product development will know that what people say they want is often very different from what they really want. And what they actually need is usually different still.

Participant observation would recognise that even when you get five people in a room together, what is agreed and written down by those five still stands a significant chance of not reflectng reality. The only way to really build software is for the builders to really know how the users will use the product. The only way to do this is to walk a mile in their shoes. Or, as a participant observer would say, to live with them for 18 months before you start writing a line of code.

Going Home

I am still struck by how many of my peers (mid 30s, poss married or in an equivalent relationship, maybe kids) still talk about “going home” to mean going to their parents’ house.

Maybe people have an idea of what “a home” looks like and compared to this the flat they own in Zone 2 seems like a stopgap measure.

Or maybe you can never really feel “at home” in a big city, unless you were born and bred there.

Or maybe it’s just a London thing.

We’re all geeks now

Over Christmas with my parents, cousins etc.

• My uncle talking about whether to buy a new computer now or to wait until Vista comes out (ok, to be fair, he called it “Some new program from Microsoft”, but even so)
• My dad talking about using eBay (hell, I hadn’t even used eBay until two days ago and I’m the one working in the online auction industry)
• My mum talking about her recent installation of broadband and Skype
• Debating with my aunt the merits or not of printing out your digital photos

Unthinkable even 2 years ago.

At the same time that technological evolution (not just Web 2.0 but even such prosaic things as Windows and USB cables) makes technology easier to use, more people use it and therefore their level of ability with it increases. A virtuous or vicious circle, depending on your standpoint.

The same goes in the workplace. Ten years ago people outside the IT department didn’t really want to know what the geeks did. Nowadays it’s often the people outside the IT department are the ones with the coolest ideas of how to use new technology to do business better. I have a theory that as much, if not more, technical innovation comes via the CEO’s kids than out of the IT department.

It’s a real challenge for Technology Leaders if they are to avoid irrelevance over the next few years. And I haven’t heard anyone with the answer yet. Though, again, I have a theory that the following will help:

1. The IT department must always be aware that what was right last year probably won’t be right this year
2. It’s pointless trying to stem the tide of future innovation just because it is inelegant or a potential security breach or because you didn’t think of it

Technologists as Translators

Conversation between mother and wife the other day about mother’s family tree went something like:

Mother (pointing at marks on a piece of A3): Look, A and B are first cousins which makes C and D second cousins but because of the age difference between A and B, D is about the same age as E who is F’s niece.

Wife (perplexed by the complexity): There must be a computer program to make all this easier.

Mother: Are you really sure that there is a computer program that knows all about my family tree?

Wife (blank look): ..?

At which point I step in to translate between the two, explaining to Mother that when Wife is talking about a computer program to more clearly draw the relationships between people (assuming you’ve typed in all the data first) and to Wife that Mother is most interested in the data rather than how it is displayed.

This is the same misunderstanding that occurs every day between users and techies when delivering technology in business. When a user talks about a “customer database” they are thinking of the data; when a techie talks about a “customer database” they are thinking of relationships between tables.

In 2006 the key challenge for anyone who wants to succeed in delivering technology is not simply understanding the technology, nor even understanding the business, but something more subtle: being able to translate between the two.

Incidentally, are there any good family tree programs out there that my mum could play with?