Kirill So

A fun watch about second-brain cringe lords claiming they woke up to Claude linking their thoughts and writing 6-8k word class articles 🙃 As a person who sadly wasted hours on this. Both painful and fun...

Looking back, after exiting the tunnel, it's crazy to realize how much influence I actually had on "my decisions".

- Many years later, move to a different country cause someone took a chance on you that took 5 minutes. Literaly 5 minutes. - Meet your spouse in that country - Thank that someone every year. So I'll leave with a wonderful quote from a great song: "Go easy kid, it's only rock and roll!"

Looking back, after exiting the tunnel, it's crazy to realize how much influence I actually had on "my decisions".

The other side of the tunnel is completely without "I DID" - Parents immigrating - Falling behind in academia due to disruption, language barrier, etc. - Starting work early due to financial scarcity - Doing odd gigs

Looking back, after exiting the tunnel, it's crazy to realize how much influence I actually had on "my decisions".

Food for thought... Startup methodologies over the last 25 years have produced zero measurable improvement in startup survival rates. Lean Startup Customer Development Business Model Canvas Design Thinking *BLS data flat curves since 1995. Crunchbase shows declining seed-to-Series-A conversion.

I don't know what the answer is, but for your own sake, I hope you are pondering too.

Food for thought... Startup methodologies over the last 25 years have produced zero measurable improvement in startup survival rates. Lean Startup Customer Development Business Model Canvas Design Thinking *BLS data flat curves since 1995. Crunchbase shows declining seed-to-Series-A conversion.

Of course, the easy and lame answer here is to create more moats, distribution, proprietary data flywheels, complex integrations, regulatory advantages, yadda yadda yadda. Stuff you've seen on LinkedIn 500x. And it's not wrong, but it's too obvious. Too generic.

Food for thought... Startup methodologies over the last 25 years have produced zero measurable improvement in startup survival rates. Lean Startup Customer Development Business Model Canvas Design Thinking *BLS data flat curves since 1995. Crunchbase shows declining seed-to-Series-A conversion.

1️⃣ Nobody runs rigorous experiments. Pundits sold books, accelerators collected returns, and academics faced bad incentives. 2️⃣ The methods formalized the obvious. Founders already talked to customers before anyone branded it. 3️⃣ The methods are self-defeating. That's the most interesting part.

680 pages, $0/month hosting. Cloudflare is pretty awesome.

I am curious how far I can push it with pSEO, AEO & GEO whilst trying to create a deterministic way to score for global professions, not only in Singapore.

680 pages, $0/month hosting. Cloudflare is pretty awesome.

Last week's numbers: 3.6K unique visitors 90K requests 2 GB served Cloudflare cached 40% of it without me configuring anything. The whole thing costs me literally nothing to run.

680 pages, $0/month hosting. Cloudflare is pretty awesome.

Some time ago, I bumped into an article called: "Risk, luck, and imitation. Did Elon Musk get lucky?" by Luca Dellana The main point goes something like this:

In 10 parallel universes, how many times does this strategy work? Are you okay with the full range of outcomes, not just the best one? As Marc puts it: "Success that costs you everything is too expensive."

Some time ago, I bumped into an article called: "Risk, luck, and imitation. Did Elon Musk get lucky?" by Luca Dellana The main point goes something like this:

This doesn't mean avoid all risk. It means acknowledging the trade-off. Competing for first place often uses strategies that reduce your average outcome.

Some time ago, I bumped into an article called: "Risk, luck, and imitation. Did Elon Musk get lucky?" by Luca Dellana The main point goes something like this:

Extreme outcomes require extreme variance. To compete for #1, you take risks that increase your best outcome but reduce your average outcome. I love this thought experiment based on the multiverse. 500 entrepreneurs take extreme risks. 500 take moderate risks.

Some time ago, I bumped into an article called: "Risk, luck, and imitation. Did Elon Musk get lucky?" by Luca Dellana The main point goes something like this:

Extreme risk strategies optimize for the best outcome in one universe. But they tank your average outcome across all possible universes. Investors optimise for the portfolio, whilst you are one-shotting it. Investor win condition -> 1 out of 10 companies returns 100x

I built an open-source site that maps how AI is changing work in Singapore. It covers 562 occupations and 88 estimated modern roles. Some jobs are exposed. Some are protected. Some are under pressure and still in demand.

Special thanks to Singapore @mom_sg, Data.gov.sg for the data sets, and @openai codex & @anthropic @claude for the help with modeling and frontend, respectively.

I built an open-source site that maps how AI is changing work in Singapore. It covers 562 occupations and 88 estimated modern roles. Some jobs are exposed. Some are protected. Some are under pressure and still in demand.

The site is more than a scorecard. You can: • search occupations • compare jobs side by side • browse estimated modern roles • see labour reality separately from structural pressure • inspect offset potential and transition support • read the methodology and download the data

I built an open-source site that maps how AI is changing work in Singapore. It covers 562 occupations and 88 estimated modern roles. Some jobs are exposed. Some are protected. Some are under pressure and still in demand.

A few numbers from the current release: 27% of occupations land in the high or very high pressure bands. About SGD 56B in estimated annual wages sits inside those occupations. 80 occupations show official demand signals overall. 23 high-pressure occupations still show those demand signals.

I built an open-source site that maps how AI is changing work in Singapore. It covers 562 occupations and 88 estimated modern roles. Some jobs are exposed. Some are protected. Some are under pressure and still in demand.

I've been silently enjoying various food videos from Japan and Korea. Just look at these bad boys! There seems to be an entire world of people just shooting how various restos do local burgers, egg balls, and cool local twists on cuisine.