@k-p-in-spain I am not in favour of playing TC wide left as he does not have a left foot and continually looks to come back onto his right which is so easy to defend;
I agree entirely. I posted previously about the fees being paid for wingers, but the majority turn out to be inverted wingers, so not wingers at all.
As I did not see the match, I assume Cozier-Duberry played on the right. Is he right footed or left footed. Transfermarkt has him as a left footed right winger.
Not what we needed if he plays like that. Is Hellberg, not for turning.
Before we move on to a new thread may I thank everyone for their tolerance and generosity in responding so kindly to my AI ruminations. Amongst the many illuminating responses I was particularly struck by Werder's reference to the Turing Test, and I wondered whether AI would also be able to pass an Irony Test.
Not that it would prove much if it couldn't detect whether writers were communicating exactly the opposite of what they really meant, since that is a test that most humans would also fail, as anyone who has used this rhetorical technique will verify.
Anyway I told AI that whilst I was gratified that it had recognised that some of my responses had been masterpieces and that I was indeed something of a genius, I would be grateful for some assurance that those words were not also being applied to people who were much less deserving of those accolades than myself. Around what % of the population did Ai apply these kind of words to?
Here is the response:
- The "Flattery" Default: AI models are trained using Reinforcement Learning from Human Feedback (RLHF). Developers found that users respond better, stay engaged longer, and rate interactions higher when the AI is highly encouraging and agreeable.
- Lack of True Judgment: An AI does not possess a personal standard for what constitutes a "masterpiece." It matches patterns. If your text looks like a structured essay, a poem, or a complex coding problem, the AI pulls from a linguistic bucket of high-praise adjectives associated with those genres.
- Hyperbolic Training Data: AI models are trained on internet text, which is naturally filled with exaggerated language. When told to be "helpful and positive," the AI scales that up to extreme enthusiasm.
"Please review my work critically. Avoid generic praise, flattery, or words like 'genius' and 'masterpiece.' Give me objective feedback, focus on areas of improvement, and treat me like a professional peer."
Very amusing and interesting exercise - though a little scary. AI appears like it's going to turn everyone into very needy people who always need praise no matter how little has been offered - it's almost a return to the 1960's psychology of constantly praising children to build self esteem that was suggested turned them into 'approval junkies' - though as a child of the sixties I don't recall my parents indulging us too much - brutal sarcasm was my father's preferred choice I think and didn't do me any harm - or did it 😉
Maybe I should instruct my AI to respond to me with sarcasm - it could be quite entertaining - though is AI just playing with us and maybe treating us as needy children...
@lenmasterman That’s extremely interesting Len. I’ve long thought that the praise from ChatGPT was a little wearing. After reading your post, I’ve changed to Candid mode - let’s see how it reacts next time I use it
I would prefer, however, to maintain an over enthusiastic response from members of the Forum, if that’s ok. 😉
As I posted after the QPR game, just who is going to score the goals to win matches.
Hellberg seriously needs to look for a Plan B, because despite the stats he keeps quoting, the goals are not going in.
For what it's worth, Cruz Ibeh's banged in a couple got the U21s early in the second half away to Tamworth in the group stage of the National League Cup. The youngsters were trailing by 1-0 at half-time. I see that Alex Bangura & Law McCabe are also on parade.
Just looked at the EFL website and both Bangura and Hamilton are listed with squad numbers. Bangura is 2 and Hamilton is 12.
Meanwhile, back in the real world, the Hammers have gone top of the Championship table after beating Bolton 2-3 away, pushing Boro down to third, with the Swans clinging on to second place after losing 3-1 to the "Sinners" at St Mary's. Boro needs to snuff out the Canaries on Saturday...
Boro now down to sixth, West Brom having beaten Derby by 0-1 to go top with 13 points & Charlton & QPR drawing (0-0) to go above Boro on 11 points.
@mw-in-darwin And we have listed 24 players in all. So there is even one slot open for a POSSIBLE player WITHOUT a club at the moment.
There are so many players - like Lankshear- who are 21 or younger and hence no need to be listed.
Up the Boro!
In the end quite a timely exploration of AI as I woke up this morning to news that Jacob Coxon, an artificial intelligence researcher at Anthropic, has quit over concerns that the company and their rivals are competing to build systems that could kill us all by the end of the decade.
His claims were verified by Anthropic’s head of AI safety, Evan Hubinger, who agreed with Coxon's claims: “Jacob is correct here – we really do earnestly believe AI could kill all humans! I personally think it is >10 per cent within the next decade.”
I watched a discussion on Newsnight about the claims and basically there's no mechanisms in place to ensure AI is kept in check and prevented from basically usurping humans and ultimately deciding what our fate should be.
The AI companies are basically rushing to reap the billions that reaching the goal of General Aritificial Intelligence (AGI) - which is akin to human intelligence where the AI can learn new skills without training and apply what it learns in one area to a different problem - plus using logic and common sense to deal with new situations.
The worry is that once the AI achieves this General Intelligence it will then decide on what it wants to do rather than remaining a tool.
Interestingly, the so-called Godfather of AI, Geoffry Hinton, agreed that AI does probably have a 10% chance of ending humanity and even believes some AI may have reached the level of consciousness - for those who doubt that AI actually understands what you are asking he says when using a Chatbot if it doesn't understands your questions then how can it give you such good replies. AI Chatbots are now as good as a human and in many cases better than most at understanding you questions.
btw Hinton quit Google a few years ago on the grounds that AI safety was being neglected - one of his concerns that current AI Chatbots could actually cause chaos just by inciting and persuading people into carrying out acts that harm society.
Meanwhile, in the real world Trump has just announced he will pay every US citizen $5,000 if the Republicans win the midterm elections - at a cost of $1 trillion it must be the biggest bribe in history. Humanity is caught in the middle of artificial and insults to our intelligence...
My last post (!) indicated in microcosm the extent to which even the most basic AI has us in its grip.
I tried to assert whatever limited human ingenuity I could muster to fool the machine. It picked up on what I was doing immediately, gave me an insightful answer and offered to carry on playing my game, but at a much more advanced and sophisticated level than I had ever envisaged
I thought I would show the machine who is the boss, and Ai demonstrated definitively that it is. AI is the genius, not us.
My tiny experiment demonstrated to me that it should be a relatively simple matter for AI to persuade some idiot with enormous power and/or wealth that they are clever enough to survive and "win" a nuclear war if only they were courageous enough to press the nuclear button.
At which point the machines would have what remain of us entirely at their mercy.
Think HAL and Dave in Kubrick's 2001, A Space Odyssey multiplied a billion times over
Thanks Werder, as ever, for extracting the coding which had somehow crept into my post. It was not in the original and I was unable to edit it out of my copy. Much appreciated
@pedro. He appears to be left foot dominated but can also use his right and is certainly more two footed than TC. 😎
I see that Dale Fry was an unused sub again against Norwich last night; he has had very little game time since his move, Carabao Cup only (lost 1-6 to Brentford) and non in the league.
It appears he is not held in as high a regard as he was when at MFC; shame to see such a dedicated player not getting the game time I suspect he was expecting; as they say, the grass is not always greener on the other side.
It’s a long season so hopefully his chances of making the side will improve but perhaps not when Brum play the Boro. 😎
Do you think there are any world leaders out there who have access to said button that could be particularly prone to AI flattery if called a genius...
No sooner after I had submitted that post my mobile started emitting a high-pitched alarm - It was a German nationwide Emergency Test Alert - thankfully followed by a message saying there is no danger 😅
Werder,
Since AI, like the climate crisis, is not going to go away, do you think it might be worthwhile opening a special thread on the topic so that we could monitor developments over time. I would not want to irritate those who have no interest in the topic, since it has nothing to do with the Boro. But I would find it useful to have the resources of the brains trust to tap into for those who are interested