I'm not a historian. I didn't come to this through academic study or political theory. I came to it through a TV show — The Man in the High Castle on Amazon Prime, which imagines an America where the Axis powers won World War II. It unsettled me enough to start asking real questions. What actually happened in Nazi Germany? What programmes existed? How did ordinary people allow it? And — the question I couldn't shake — are we, right now, closer to that world than we think? What followed was a long, uncomfortable rabbit hole through history, AI, power, and human nature. This is what I found.

“Every atrocity in history followed the same first step — making human dignity conditional. We cannot afford to repeat that in the age of AI.”

From the Aktion T4 programme that murdered 300,000 disabled people under medical authority, to today's algorithmic systems denying welfare and housing to the most vulnerable — the pattern is unchanged. Power concentrates. Accountability diffuses. And the people with least voice bear the greatest cost.

The AI revolution is not a new story. It is the oldest story, running on faster infrastructure. Manufactured trust replacing genuine care. Bureaucratic distance laundering moral responsibility. The access gap widening between those who shape these systems and those shaped by them.

But moral progress is real — slavery abolished, democratic accountability expanded, women's rights advanced — not through perfect systems, but through people who refused to accept the current dispensation as permanent.

The foundation isn't a new law or a smarter algorithm. It is something every tradition across every century has independently arrived at: human dignity is unconditional, irreducible, and non-negotiable. Not earned. Not conditional on productivity, race, or utility. Simply held — by every person, without exception.

Applied to AI, this means one thing practically: every system that makes consequential decisions about human lives must be able to see the individual, not just the category. Must be accountable to a human face. Must distribute its benefits toward those with least power, not most.

But here is where I have to be honest with myself — and where I think you do too. It is easy to write about the failures of systems, institutions, and leaders. It is much harder to admit that I have also been on the wrong side of this. That distance — between me and another human face — has sometimes made it easier to justify a decision that served my convenience over their dignity. A dismissal I dressed up as efficiency. A boundary I called principle but was really just comfort. We are all capable of this. The moral high ground is not a place any of us permanently occupy. The finger we point outward must, honestly and regularly, turn back toward ourselves.

Your part in this

Ask one question of every system that affects you — and every decision you make that affects others: whose dignity does this serve — and whose does it cost? You don't need expertise to ask it. You just need the courage not to stop asking until you get an honest answer. Starting with yourself.

#Gemini #GoogleAI #Multimodal #AudioAnalysis #AI #BerniePng #TechTrends #Productivity #LLM #Gemini15Pro #AITesting #FutureTech #VoiceAI #MachineLearning #Innovation

Without the huge repository of human creation through the centuries (the inventors, the artists, the philosophers), AI wouldn’t have a chance to synthesize our knowledge to help the rest of us (the common folk).

AI just made knowledge and creativity more accessible.

But we still need the human geniuses among us to create and innovate. Our imagination needs to prevail for the future of humanity.

That is our value.

Only then, can we work side by side, and not behind, AI.

On a side note, we should re-think how to protect human creativity. Current copyrights and trademark laws need to be revamped.

What do you think? (I'm asking the humans) :)

#AI #HumanAICollaboration #FutureOfWork #HumanCenteredAI #ArtificialIntelligence #Innovation #DigitalTransformation #HumanTouch #AITrends #BerniePng

I'm (still) loving every bit of it, and my brain is on fire. Check in on me in 2 months, I might have a different story to tell. Might be fried by then.

Three key mini inflections took place in my life with AI this month:

• 1st week of March

  • The Deliveroo exit from Singapore. I did a symbolic mini tour around almost all the zones, March 1 – 4. It was nice.
  • My attempt to get hired as an apprentice with AI Singapore. On March 4 they sent out the technical assessment to everyone who applied for the apprenticeship program, giving us 6 days to work on it then submit. Wish me luck.

• 2nd week of March

  • I met up with a few of my ex-colleagues. We talked about the future and our plans. It was nice to catch up with them.

• 3rd week of March

  • With the GTC 2026 and SXSW keynotes combines, I finally understood where we are heading, and really gave me a lot to think about, mainly how what value can I offer in this new world, and what gaps I have.
  • I paid for Claude for a month, but have not explored Cowork or Code in depth. I'm still able to build lots of stuff just using the chat.
  • My class assignment for this module 2 involves Big Data. I ingested 23 million rows from a public dataset and did analysis on it. It was fun.

At some point in the keynote Jensen said that in the future we would not just be paid in money, aka salary, but also, with tokens. I believe him. But what does it mean for us, the common folk? He talked about AI factories and showing off all his highly efficient racks with awesome “tokens/watt” servers, and how everything will be opensource. It's like “here's all the free chocolate and candy you can eat, but you have to pay for the teeth.”

Welcome to the new world.

#AI #LLM #FutureofWork #JensenHuang #OpEx #TechnologyTrends

Here's the prompt I used.

Hi Claude you are an environmental consultant tasked to study how air quality of a country is affected by the activities within in, and you are doing a big data analysis to understand, rank and visualize the top 5 industry polluters. You are also to compare these polluters across different countries and show how, if any, what quantifiable improvements made by each country and their results. Guide me through step by step using Google's Cloud Platform's Big Query to ingest a free dataset through API or other real-time data ingestion, preferably streaming, then use a suitable data warehouse to store the data and then do ELT pipelines, then create some prediction models and generate reports through the transformed data. Your final output should be a modern dashboard that can be presented to the United Nations meeting to share the challenges of balancing advancement vs air quality and how poor business decisions can lead to bad air quality and eventually poor health of the people.

So now, it's day 2 and we are working on creating a workflow of using Docker to run Spark to get Air Quality data through API. I'll keep you updated on what happens.

#AirQuality #DataScience #BigQuery #MachineLearning #EnvironmentalData #GoogleCloud #ClaudeAI #DataAnalysis #Sustainability #TechJourney

As I'm going through the free Stanford Online course CME295, I thought about Natural Language Processing (NLP) and how we currently tokenize English text to feed into LLMs. If we transform words into numbers, how do we add in culture nuances, slangs, etc that are unique to a country so that we can create a sort of naturalized citizen? What is Singapore currently doing?

So I asked Gemini for some thoughts: so if I am a citizen of a country who wants to help build a realistic LLM for my country where we have 4 languages and we want to not only create a balanced LLM, but also to incorporate our country's culture and slangs into it so that the LLM performs like a "fellow citizen" and knows all our nuances and history. How should I start?

Gemini's answer: “This is a noble and technically fascinating mission. Building a “Sovereign LLM” is the move from a model that knows about your country to a model that is of your country.

Be careful with “Data Poisoning.” If you only scrape one political side or one ethnic group's slang, your “Fellow Citizen” AI might end up being biased or exclusionary. Balance is your hardest job.”

It then went on to suggest some open-source models to start with and actionable next steps.

[NOTE] I'm using this blog article to document the start of my journey. It will be a multi-year quest of mine. I know Singapore is already embarking on building her own LLM or LLMs, but wouldn't it be great if I can create my very own Singaporean friend, a super knowledgeable, kind and caring one at that.

Cool idea?

#ai #LLM #NLP #SovereignAI #Singapore #Singlish #AISingapore #CME295 #LearningInPublic

This morning I felt a sense of emptiness.

There has been some collateral damage as my Bug Farm was not tended to diligently, causing the mealworm lifecycles to overlap in the same bins. If you don't know what I'm talking about, it's ok.

So, today I'm going to spend most of the day tending to my worms, then for Claude I'm going to test out some stuff:-

• Learning more about skills and structuring my coding workflows to utilize this feature
• Changing Claude's setting to “brutally honest”
• I have a mini project I codenamed “Under the Hood”, which I will start on and will update you soon

Oh yes, editing my video for Bernie Studio Delivery

Have a great week ahead!

#reflection #PositiveThinking

I recently had a realization while watching Dispatches from Elsewhere: staying quiet is a form of selfishness. In the show, Fredwynn tells Peter that by not sharing his perspective, he’s withholding something valuable from the group. That hit home. For too long, I’ve been a quiet solo-builder, but as I pursue an apprenticeship with AI Singapore (AISG), I’ve realized that my mission, and the knowledge I gain, belongs to the community.

Singapore has no natural resources; we only have our people. I believe AI is our unique opportunity to transcend our physical limits and export local talent to the world. To get there, I’m attacking this from three angles: building real-world solutions, earning technical certifications, and finally, refusing to be 'selfish' by sharing every step of this journey in public.

• First I will just solo-build, or team-build (I don't have many friends to start with actually), help my friends solve their issues, learning and building at the same time.
• Secondly, I will build my credibility by getting certifications, skills and knowledge
• And third, I will resolve to share more about my journey, because I recently realized that if I don't share, I'm being selfish with my knowledge. So ... I will share more, here on my blog, and also look out for upcoming videos on YouTube.

If I don't get the apprenticeship I will still keep building.

Thanks for reading.

#reflection #ai #apprenticeship #AISingapore #AISG #SingaporeTech #SoloBuilder #AIGovernance #SmartNation #CareerPivot #LearningInPublic