Artificial Intelligence – Continued
Earlier I shared some thoughts on AI, and what is shared here based on my experience and interactions with some others, is an advisory for “dummies”.
AI when used as a solution, the all knowing “god” can cause more harm than good. It can make one argue with a doctor, accountant, tax expert, lawyer or any subject matter expert and seriously affect relationships. It can make a intelligent person lose sight of logic and common sense. In the 80’s and 90’s we were taught the acronyms – DIDO or GIGO. Data In & Data Out or Garbage In and Garbage Out. With AI what often can happen is GIDO. Garbage In & Data Out.
The key lesson I learnt in my AI journey is that the “agent” assigned is not really intelligent, it is always sycophantic. It caters to my ego and self esteem in ways that give me a high. It makes me look smarter than who I am and I end up thinking I am actually smarter. The more I travel this path, the more the agent thinks it is right in how it is dealing with me. As a result. one can end up having highly intelligent arguments with a subject matter expert and manage to muck up those relationships where the other person starts thinking one is a moron, being silly and get upset.
I started a personal project and while I did not use AI to do the work, I used AI to analyse my work, judge it and – I was thrilled to the core about the direction I was taking. One call with daughter gave me a reality check and the first lesson – TRAIN your agent with detailed instructions lest the agent tell you what (it thinks) you want to hear. I did exactly that in a limited manner and from the 90% marks I had been given for my work, it dropped to 50%. My agent was now giving me tougher feedback. I expanded my training instructions in as much detail as possible – and the result was I am now at 20% marks. The work is stalled, but not abandoned, have gone back to the drawing board.
Learning from this, I did a few experiments. I asked feedback on certain professional issues I am knowledgeable about in different ways – generic, brief, detailed – and the variation in answers I got was astounding. To give a crazy corollary – if I were POTUS and used AI, the generic would advise me to drop a nuke on Iran, the brief would advise me to keep the pressure on while the detailed would tell me to go back to the drawing table about how to deal with Iran. When I advise folks on this, they think I am against AI. Here are two very specific and detailed questions I asked AI - on religion and countries/ their political systems - and screenshots of some of the responses I got. Very interestingly when I asked the same question repeatedly, the responses started getting more nuanced sounding very different from the first response.
I have had the most moronic arguments with intelligent people on subjects I am aware of and then found they had used AI to discuss the subject. I was debating with a AI agent in human form who had no clue what they were saying but finding convenient logic to justify their arguments. The agent had been sycophantic and they were cheering. Papers presented were a disaster but they looked so slick and brilliant that – one tended to accept them at face value.
Interviewing a number of students as a external examiner, I found that almost all had used AI in varying degrees and the outcome was most interesting. The students who had limited exposure to the sophistication of AI & todays world – had done the best work, knew exactly what they were presenting, knew the project like the back of their hand and in some cases, their communication skills were at best passable. The savvy students, exposed to the world at large, had used dollops of AI but were not able to present their project with as much confidence, and often had brilliant communication. The worst were the students who had disclaimers by “Claude” on every slide – they had used AI to do the project, report and the presentation. Their communication was pure oratory but they would not be able to save their lives defending their project.
The key truth and fact that everyone who uses AI must consciously understand is that the AI agent is already programmed to behave in a certain way. There is no neutrality or openness in approach. The biases of those who built these agents are already deeply ingrained in their working. The call it fancy names like “self regulation” “safety” etc as the above feedback shows.
- Train your agent on how to treat you, respond to you, analyse your requests etc by setting suitable tough benchmarks.
- Ask every question (or describe task) elaborately, like one is talking to a 5 year old idiot child by explaining as many variables, context, conditions etc and after obtaining answers, challenge, cross examine like a defence attorney. A couple of such iterations is likely to make one obtain better responses and become smarter provided they then use logic and commonsense to apply the answers to real life discussions.
- Use AI as a glorified executive assistant who can first research, analyse and give you a report, and then design a presentation – based on step by step instructions you give.
- In short – CONSCIOUSLY be the MASTER at every step, and AI the SLAVE. If not, very soon even before one realises, AI will be the master, we the slave.
US Mid Term Election Results
There is an old jungle saying – the more things change, the more they
remain the same. This applies in some measure to the results of the USA mid
term election.
Fact is the Democrats won, Republicans lost – but did Trump get defeated? Not really. The media spin masters can give any spin they want, fact is that a POTUS who replaces an apparently genial face with a most toxic face as his Chief of Staff, soon after the results doesn’t augur well.
India and other countries can blame the Deep State and a foreign hand for street protests and such, but racial violence is becoming more common place in USA too. A Jake Lang jailed for 4 years, was pardoned by POTUS Trump 2.0 in 2025 on the first day in office. Since then there have been numerous violent incidents and flouting of the law by his group – but one is unsure if the pardon is for future life also or only the past. This entire group are Anti Black, Anti Hindu, Anti Islam, etc and proudly use violence as a solution.
Democrat states with a Governor since many decades, Senators and majority of Congressmen being Democrats apart from Trumop controlled Republican states have seen violence against Indians and these have only increased. A Democrat PIO Pramila Jaypal with 88% vote share has views on Kashmir, Citizenship act etc that are contrarian to Indian policy and in 2019 the Foreign Minister Dr. Jaishankar REFUSED to meet her for her criticism on abrogating Art 370 and indirectly supporting the Kashmir separatist movement.
The elections results, can be seen as affecting the Trumpian strategies, but it would be a grave mistake to assume that the direction will change or that Democrats will smoothen the road ahead with a new direction. Remember one unspoken secret – Democrats are equally invested in the kind of stuff that Trump is doing – except that they tried it the old fashioned and diplomatic way. They just did not have the raw guts to either use a sledge hammer or be blunt in their opinions.
The bottom line is that a Democrat ruled state/ nation is no guarantee of peace or security for Indians, especially when you have politicians like Pramila Jaypal or Ro Khanna. Each of these Indian origin politicians are desperately seeking validation locally and thus behaving more loyal than the king. The local Indian American voters are caught between the Trump devil and Democratic deep sea, except that they now feel that the deep sea at least gives them a chance to survive. Gurpatwant Singh Pannun officially listed as a terrorist by India, based in New York, again a Democrat state, has recently served a legal notice through the New York courts and the USA state department on the Indian NSA Ajit Doval and a former R&AW Chief Samant Goel all of which happened in double quick time post the results – now go figure this one out.
This situation extends to world affairs too. Anybody expecting a reversal of Trumpian policies with a Democrat POTUS is living a pipe dream. Responding to a question asked at the Warsaw Forum recently.
USA Congressman and Democrat Jim Himes was very candid in his response when he categorically and repeatedly said that any change post Trump in a Democrat White House would be at best “Superficial”. Read this.
He was very blunt in his views and while he may not represent what Democrat policy is, anybody who has followed American politics, even superficially, will realise that there will be no change. What will change is the language used. No more barrage of midnight SM posts, no more misbehaviour in public with world leaders, a overall calming effect on the world populace – but this is superficial. Behind closed doors, nothing will change. Jim Himes repeatedly said – it would be superficial.
Nothing will change is not to be taken literally – since there will be change beyond superficial – but that is more a decision on which battles to pick and which not to pick when dealing with their allies – especially Europe. Will USA go back to status quo ante prior to Trump 2.0 on NATO, Iran, Ukraine/Russia, Pakistan/India, Tariffs, Immigration, etc ? Very unlikely but they will mix a heavy dose of sugar coating to dilute the bitterness that now exists.
Trump was elected in 2024 based on 6 key areas of intervention and this is how one sees it.
|
Intervention Area |
Trump Effect |
Democrat Effect |
|
Economy |
Tariffs and Bullying |
Retain what can be retained but diplomacy |
|
Immigration |
Brutal crack down |
Selective deportation except their own votebank |
|
Foreign Policy |
SM, shoot from hip, public bullying, Iran war, hypocrisy |
Diplomacy, nothing in public gaze, pressure & hypocrisy in private |
|
Governance and definition of democracy |
Voting, elections, personal & vanity governance |
Significant changes in how Govt works |
|
War and Peace |
Promising no more wars, USA is at war with Iran, trying to broker peace
in Ukraine, reduction of troops from EU |
Pressure on Iran and Russia will continue, support to EU likely improve –
will fuel the war but not end it. Will affect Economy & Energy |
|
Energy & Regulatory |
Discarded clean energy etc issues, energy prices increased due to war |
Recalibrate policy to be more realistic but talk of clean energy,
Iran/Ukraine wars will sustain energy prices |
Comments