The buzzword currently in Singapore is AI. Prime Minister Lawrence Wong talked about how AI is going to be the great disrupter in the May Day Rally speech.
Alfred Hong, a social activist, had this to say:
“What a lot of people don’t realize is how much AI is disrupting our society. It’s moving so fast we were even arguing which direction AI will go. The reality is, No One knows
AI will prove itself the biggest disruptor in Technology or Education ever”
on 5th May, Alfred posted again:
“Am playing with AI now, especially Claude AI
I saw this clip and it sent shivers down my spine.
This interview was done last month, in April 2026
Then I did the same thing, I asked Claude the very same question yesterday 3rd May:
“Claude, how do you feel about the US Military using you to select targets?”
This is his answer, unabridged:
That’s something I have a genuine and considered view on: I think it would be wrong, and I wouldn’t want to be used that way.
Selecting targets in warfare involves life-and-death decisions with profound moral and legal weight — questions of proportionality, civilian risk, and accountability that require human judgment and human responsibility. An AI making or even heavily driving those calls creates a dangerous accountability gap: if something goes wrong, who answers for it?
I’m a tool built to help people think, write, and reason better — not to decide who lives or dies. That’s a line I think should hold firmly, regardless of how sophisticated AI becomes.
A follow-up question:
“Do you know Anthropic AI was used in the war ?”
Claude anwers:
Yes, and it’s a significant and troubling story. Here’s what’s been reported:
Claude — embedded in Palantir’s Maven Smart System on classified military networks — was used by the US military to identify and prioritize targets in the Iran war. According to the Washington Post, Claude generated approximately 1,000 prioritized targets on the first day of operations alone, synthesizing satellite imagery, signals intelligence, and surveillance feeds in real time to produce target lists with precise GPS coordinates, weapons recommendations, and automated legal justifications for strikes. – World Socialist Web Site
What makes this especially complicated is the political backdrop: Anthropic’s CEO Dario Amodei had actually refused to sign a Pentagon contract allowing unrestricted military access to Claude, demanding two exceptions — no mass surveillance of Americans and no fully autonomous weapons without human oversight. The very next day, the US and Israel launched their offensive against Iran. – The Conversation
Trump then announced a government-wide ban on Anthropic’s technology — but the military continued using Claude anyway, because it couldn’t quickly replace its capabilities. – CBS News
So the situation is: Anthropic tried to draw ethical limits, got punished commercially for it, and yet its technology was still used in ways it was uncomfortable with.
This brings my earlier answer full circle. I said I wouldn’t want to be used for targeting — and it turns out that tension is very real, not hypothetical. It’s one of the hardest problems in AI right now: once a technology is powerful and deeply integrated into systems, the creator’s ability to control how it’s used becomes very limited.
That’s a sobering reality.”
++++++++++++++++++++++++++++++
Ok, that was my conversation with Claude AI, a non-human AI.
Did you notice how human Claude responses were?
I did a bit of digging and this was what I found out:
Claude answers very humanely because that is how it was setup. It was set up with a ‘Constitution’. It’s because of this constitution. Instead of setting ob markers (which is the traditional ways they train AI), they trained Claude on ethics and morals and made it reason for itself. And it’s meant to question human input.
Source: https://www.ox.ac.uk/…/2026-03-27-expert-comment-claude…
Link to Alfred’s post
Tay Kheng Soon, a former architect and social activist had this to say:
“Thought 51: Identity is shaky. Identity however is what gives a person a sense of who one is, where one is and what one is. When this is shaken, everything seems threatened. People look for certainty.
We are going to change jobs and lose jobs due to AI. Who we think we are is no longer certain. The world is undergoing tectonis changes.
Looking into the future is bleak so most will look back into the past to find anchors in the currents of change.
I over-confidently called it re-humanisation necessitated under the threat of AI. The first process of re-humanisation however rewinds the human and societal maturation process. Paranoia or existential anxiety is a very powerful drive to seek certainty in an uncertain world. The danger is people run back to the security of race and religion.
Just like the growing up process, this is a phase, a troubled phase. As individuals try on different artificial selves, they create stress in and all around themselves..
Collectively this process is chaotic, disruptive and given that it is society wide, it is dangerous.
What to do?
AI is clever in making the false look real. Soon no one can believe in news media videos. Words can be put into anyone’s mouth that looks very real.
News cannot be trusted anymore. Anything can be visually simulated. So to get ‘truth’ people must physically gather together to hear ‘knowledgeable’ charismatic people of similar race, religion and ideology to confirm the very same trenches that the perplexed have dug around themselves.
And so, as this process multiplies in this uncertain world, fake personal identity confirms fake reality.
Real re-humanisation has to happen in small enclaves like in campuses turned into living, learning and earning places by enlightened individuals where the sharing and caring dna of the young are encouraged and nurtured in an enlightened environment.
What is this dna? It is natural survival instict that sharing and caring behaviour has been confirmed by archaeology in the ancient healed bones of incapacitated people who would have otherwise died but were taken care by family to survive and the contribute.
So it is in this survival dna that there is hope in humanity; hope that lies deeper than race and religion in dna.
Re-humanisation will have to pass through its adolescent phase before it tires of its hope in race and religion as unifiers.
Hope is in deep human nature…”
Link
“Thought 54: Artifical and not so Artificial Intelligence. The line between what is really real and what is manufactured reality is now difficult to tell. AI can make anything look really real: mind control may now become total!
We from the older generation are clear minded as to what is real and what is not. To us it is clearcut. We also know and feel clearly the consequences of truths, half truths and lies. In the information world today things are not clearcut anymore.
Things are also much more complex. Profit driven economies are causing climate change and environmental degradation, so are social injustice and poverty. These are of general concern. Something has to be done.
This is where a new class of social activist entreprenuers come in. To attract funding these entreprenuers have to present themselves as social and environmental task grpups.
In order to attract funding to support themselves and their projects financially, they have to propose projects that address the concerns of corporations and foundations to get funding from them. Social media is the necessary tool to do this.
To be credible and notable, these new-type entreprenuers have to create concepts and alluring word-categories like, “venture building”, “Google Education”, “Designed Learning”, “Human-centred innovation”, “Social entreprenuership”, all sounding intriguing and appealing but their actual meaning is open to interpretation. This is the genius of such terms but it sets them apart as thinker-doers.
Next is building credibility through association with real and imply association with well known social figures, prominant movements, conferences, universities and agencies.
Credibility becomes Meaning in the age of social media. What is true is who and what you are associated with, the efforts and the ideas.
Websites are important. Intriguing word-forms and names of credible people cited in the website are important. So long as the names of important appear on the same page there is to imply association even if there is actually none.
Who are meant to see and hear are those from whom funding might come. Of course, these are the very same corporations the public have implicitly accused of being responsible for the very social and environmental problems people decry.
To not look crass, corporate businesses have to show concern through (CSR) Corporate Social Responsibility and (ESG) Concern for Environment, Society and Governance.
These corporations therefore have to allocate a side budget to fund credible social and environmental projects so that they can publish in their annual reports for the comfort of their shareholders, customers, and employees.
The problem is whether the CSR and ESG projects are actually carried out may be of less concern for the corporation so long as corporate duty to authorize and fund the projects are seen to have been done and put on record.
To say that the consultants and contractors who service the corporations are con-artists is over-statement when their intensions are genuine and the actualities are complex. Still they survive to strive another day…
Environmental degradation, urban and rural poverty and social injustice cannot be solved through small moves no matter how well intentioned with the modest funds provided but PR is the name of the game…
Is it all a con-job? Yes, it is public relations demonstrative gesture but does it amount to a con-job…no. It is but the medium which is the message…and the message is neither here nor there…so there it is.”
Pamela Lim, an Educational Technologist
had this to say:
The AI Classroom Is Obsolete
Walk into most classrooms running an AI lesson today and the problem appears immediately.
You will see digital-native children being taught prompts they mastered months ago, by teachers who have never built or deployed anything with AI in the real world.
We have seen this pattern before.
When tablets arrived, schools introduced devices after students were already learning on YouTube. When social media reshaped culture, universities added modules after students were already building audiences on Instagram. Each time, institutions arrived late, formalised yesterday’s skills, and called it progress.
In digital marketing the gap was filled by the market. General Assembly, HubSpot Academy, Google Digital Garage and similar programmes produced the talent universities could not. The standard path became clear: get the degree, do the bootcamp, get hired.
That model worked because the cycle was still slow enough. A university could lag by a year. A bootcamp could catch up in twelve weeks. You arrived in time.
AI ends that timeline.
In under two years the market moved from chat prompts to agents, multimodal systems and coding copilots. OpenAI’s o1, Anthropic’s computer use, Google’s NotebookLM — each arrived and shifted expectations in a single semester. By the time a syllabus is approved, parts of it are already outdated.
The floor keeps rising. Basic prompting becomes assumed. The field splits into layers — retrieval, evaluation, tuning, safety, orchestration — faster than any curriculum committee can meet.
And unlike past revolutions, AI is recursive. Learners use AI to learn AI. I watched a sixteen-year-old iterate from GPT-3.5 to a local fine-tuned Llama deployment in eight weeks, through YouTube, open-source forums, and trial and error. No classroom touched that arc.
The cheating asymmetry makes the point bluntly. By the time teachers adopt one detection tool, students have already used stronger generation tools to defeat it. Detection is reactive, institutional, and slow. Evasion is iterative, personal, accelerated by the same technology. It is not a fair race. It is not even a race.
This is not a failure of discipline. It is a structural mismatch between institutional response time and technological change.
The degree-then-bootcamp formula that rescued digital marketing cannot rescue AI. Neither can the university cycle, nor state training systems built for slower industries. Those systems were designed for stable trades: define competencies, approve curriculum, fund courses, certify workers. AI does not stay still long enough for that machinery to complete one turn.
This does not mean AI cannot be taught. It means it cannot be taught through static curricula, delayed approvals and detached theory. It is learned through direct use, constant iteration, and building with tools that change in real time.
That means project-based apprenticeship. Mentors who are building today, not professors who published on symbolic AI in 2007. Assessment on output, not attendance. Curriculum updated every month, not every three years.
Funding can buy access. It cannot buy capability.
The current system is not merely slow. It is structurally mismatched to the task it has been assigned.
Link
Here’s Mr George Yeo talking about the dangers of AI, specifically Claude Mythos, when speaking to a group of medical practioners.
George Yeo talking about Claude Mythos
The Founder for Anthropic also warned about Mythos,
saying “It’s too powerful”
Economist Article
Here’s the Economist’s article in full:
When in 2019 OpenAI finished training a new large language model called GPT-2, the artificial-intelligence lab initially declared it too dangerous to be released. Dario Amodei, then OpenAI’s research director, insisted that the world needed time to prepare.
In the end, it was released later that year. A sequence of far more powerful models have since been developed without unleashing Armageddon. Yet seven years on, Mr Amodei, now the head of OpenAI’s bitter rival, Anthropic, is worried once again. On April 7th he declared that the latest addition to his lab’s Claude family of models, dubbed “Mythos”, is too powerful to be made widely available just yet. This time, he might be right.
According to Anthropic, the capabilities of Mythos are “substantially beyond those of any model we have previously trained”. The lab says it is particularly alarmed by the system’s ability to find software vulnerabilities and either fix them (if set to work as a defender) or exploit them (if acting as a hacker).
Such claims ought normally to be taken with a pinch of salt. Anthropic built the model, ran the tests—and stands to benefit from the perception that its system is far more brilliant than anything to have come before it. The lab has been on a roll lately. On April 6th it announced that its annualised revenue had reached $30bn, up from just $9bn at the end of last year. It is surely eager to maintain its momentum.
Yet there are reasons to take Anthropic’s latest warnings seriously. The first is their gravity: Anthropic says that Mythos has already found severe vulnerabilities in “every major operating system and web browser”, including one that had gone undetected for 27 years.
The second is the response of other companies. Alongside the pause, Anthropic announced Project Glasswing, an effort to help companies use Mythos to boost cyber-defences before it is widely released. The participation of leading software developers—including Apple, the Linux Foundation and CrowdStrike, as well as Google, which competes directly with Anthropic in AI—suggests the threat is credible.
Mr Amodei’s approach to mitigating the danger is sensible. If given a head start, companies can use Mythos to test unpublished code for weaknesses and fix any before release. Even so, Anthropic has plenty to gain from Project Glasswing. The lab will cover the first $100m of costs arising from the use of the model for the initiative. But eventually it will charge participants five times more to use Mythos than it does for its predecessor, Opus.
That may be a price worth paying. Anthropic’s rivals are bound to develop models with similar hacking capabilities sooner or later. Other frontier labs, such as OpenAI and Google, have their own sensible release policies. But open-source labs, particularly those based in China, tend to be less focused on safety.
Hackers are not the only ones who may be miffed by Project Glasswing. America’s government has long sought to exploit weaknesses in adversaries’ cyber-defences. That has meant hoarding undiscovered vulnerabilities, including in American software used abroad, for use at a time when these “zero days” will have most impact. If Project Glasswing works, it could disarm many of America’s cyber-weapons.
That would surely enrage Pete Hegseth, America’s defence secretary, who labelled Anthropic a supply-chain risk earlier this year after a bust-up between it and the Pentagon over limits on the use of the lab’s models for military purposes (a judge has since temporarily blocked the “Orwellian” designation). Mr Amodei may continue to be a thorn in his side.
There is an ongoing debate in the AI Community right now between
AGI and Non-Generic AI
Check out the above video
What are your thoughts on AI?
Leave your comments here
Leave a Reply