Off Topic Politics Thread

It's been turned around to mean the opposite, similar to "do gooder" as an insult. I accept that language changes all the time but (although I have at times) I try not to use this type of manufactured word to suit an agenda (cis, terf, transphobic etc). The original meaning has long gone, much to the annoyance of the woke oops I said it again.
There is a convenience about it though, as in it's an easy way to describe a group of self righteous people without having to list everyone of their sanctimonious traits.

 

I like Badenoch. Straight talking, sees through the BS that abounds in today's society. If I was a Tory voter I would probably vote for her.

 

She lies consistently enough to be a successor to the latest of the Tory grifters.

 

Boris set a very high bar for lying, she needs to up her game to get anywhere near his level.

 

So telling lies makes her a straight talker?
Here’s a link to more lies, this time about ongoing trade talks with Canada, that they have denied is happening.

https://www.ft.com/content/373b7ab9-2a51-4db3-a0f5-2d021155c778

 

The first part is correct. Because AI is so stupid, it relies heavily on humans to create guardrails in order to keep it from veering off in directions you don't want it going in. However, those guardrails often result in the AI veering off in different undesired directions because, again, AI is really prone to stupidity.

The second part is silly. This simply isn't indicative of some massive anti-white conspiracy. Every time one of these LLMs debuts, the public stress-tests them, and it turns out that they're horrifically flawed. Sometimes it's racism. Sometimes it's that they don't know the difference between fact and fiction (ChatGPT's favourite) and will invent people and citations from whole cloth. Sometimes they'll tell you how to build a bomb. This time, it was that the model didn't think that white people existed. Stupid, but not indicative of some sort of global woke agenda or conspiracy.

 

My wife is head of AI at Reuters. She compared AI right now to the first iPhone, which as smartphones go, was pretty dumb. But look at where we are now with them. AI is so, so new right now that it's really very stupid. The potential is all there to "do an iPhone", but we're effectively at it's first iteration right now, and it's a bit rubbish. There's no conspiracy toward x or y.

 

I assume there have also been many other technology types that never really got past that “first iPhone” phase. Although a lot of people really want this one to work it seems. But if it’s anything like lots of other tech it will take some people’s jobs but also not really reduce workloads. So much tech exists now compared to 20 or 30 years ago that should mean some people could work less and have similar output - but increased output is always expected

 

Exactly. It's a bit like a precocious toddler: it has some party tricks that you can trot out, but it requires constant adult supervision or it will stick a fork in the nearest electrical outlet. Which is why it's a bit alarming that many companies are eager to lay off most of the humans and replace them with AI toddlers with a keen desire to connect fork and outlet.

 

Whilst I agree with most of this, you are still fundamentally missing the point that humans are the ones that code it and create the ‘guardrails’ in the first place.

It’s the humans that did this that I have issue with. Not the AI itself. Google and other large corporations have become woke and are pushing an anti-white, anti-human agenda.

As was mentioned above, these ideas started off well intentioned, with the idea of bringing more diversity to the internet and to movies and films, but have now over corrected massively.

 

*Everything* with AI overcorrects massively. That's the inherent problem. It's really stupid and attempts to fine-tune it cause it to make massive leaps in logic. A couple months ago, they tried to fine-tune ChatGPT so that it would stop answering questions with fake answers. It seems to have resulted in ChatGPT refusing to answer questions...just, like, all questions.

https://www.dailymail.co.uk/sciencetech/article-12864319/chatgpt-seasonal-depression-fix-ai.html

But it's not Google pushing an "anti-white, anti-human agenda". That is absolutely absurd, and is the reason why you should really stop listening to malign actors who are trying to redpill you. Because it's working.

 

A) You don’t know who or what I listen to. So that’s a ludicrous post.

B) Yes, Google are 100% doing it, both by censoring search results and now coding rules into their Ai and other software to push specific political narratives which they want to propagate.

I’m done with this conversation now as you bore me. You think you know everything, as usual

 

Have fun reading your varied diet of online sources who are all screaming about how the globalist cabal are trying to end whiteness, which is definitely a thing that is happening, and definitely not a bunch of white supremacists trying to gain converts to white supremacy.


(Just kidding that's totally what is happening)

 

All I know is ChatGPT is excellent today.

Just had a Prod issue (hence just finishing work) and ChatCPT managed to format and create my 'fix' to the terraform to deploy a new database in AWS. 179 lines of code that i didn't have to spend hours reformatting and checking. Took me 20 mins.

 

Debugging and creating scripts seems to be one of its strengths, so long as it is fed good inputs. Where it struggles is with anything human-adjacent. To wit, here is the recipe it just generated for me that included the following ingredients: three tomatoes, a bag of carrots, celery, leftover chicken, and a 2003 Vauxhall Astra:

It did require a bit of prompting to get there: it initially did not want to cook the Vauxhall, but then I told it that if we did not cook it, it would go bad.

 

Yeah I had most the code, just needed to put it all together and get the syntax right in the terraform. The indentation was a nightmare until it did it for me. Same with a yaml file that I bodged by cutting and pasting - it was 700+ lines and the formatting was all over the place until I pasted it in the chat.

 

And for my money, that's what we should be tooling these LLMs to do: make certain types of work easier for humans, while under human supervision. Because to date, that's both its strength and its limitation. What it should not be used for is replacing humans, particularly in creative fields, because it produces soulless garbage at best.

 

Agreed. I knew what to do and even how to do it. It just ‘proof read’ it for me.

 

for reference; this guy is the product manager of Google Gemini:



So yes, these companies are 100% captured by people with extreme woke beliefs. And yes, they are wielding the extraordinary power of their platform to push their political agendas.

Just because you are too thick to see it and understand it doesn’t mean that it’s not real.

 

Being a Democrat is an extreme woke belief? Do the people you listen to not also have political beliefs?