LLMs can now easily make great charts
One thing with using LLMs is that the value of knowing WHAT to do is still very important, but the LLM can fill in the “how”. A consequence is that I’m seeing better charts nowaday.
Life
Pertinent Observations.
Karthik S on startups, meditation, job hunting, and personal reflections.
Recent essays
30 of 69Performance
On Sunday, just before we left for our holiday to Sri Lanka, we started dismantling our dasara doll display.
Clothes diversity
When there is a small number of “good” brands there is a high chance that you wear the same thing as someone you meet. And that can be embarrassing at times
Leopards in motion!
Finally I saw a leopard walking in the wild today!
Comfortable weather days in India
Weather wise, Bangalore is easily the most comfortable big city in India weather-wise, but it has clearly gotten less comfortable over the years
Beckerian inversion
Maths used to be a “Beckerian discipline”, with mathematicians contributing to several other fields. Now, LLMs are proving theorems and they are absolutely not happy!
Try my new Chart Doctor
Now you can use my app to fix bad data visualisations!
Conference design
I had assumed a well designed conference is one where organisers just get out of the way and let attendees make serendipitous connections. But perhaps we need to do more to foster such connections
Studs and fighters, and the Pareto Principle
One big disagreement between studs and fighters is in regard to where on the pareto frontier one might operate. Studs take calculated risks, and can thus be more efficient. Fighters are more rigorous.
Aging societies and economic growth
Aging societies are not great for economic growth - voters prefer to focus on maintaining status quo and pensions. How will this turn out in India, given we’re growing older?
How I make presentations nowadays
I first get an LLM to make the first draft, in a markdown and a PDF. And then I just edit the markdown. And then get the LLM to knit it. Formatting and all is vibved.
Binge shopping and vibe coding
Vibe bug fixing is like binge shopping - some items / fixes make you so happy you end up not paying sufficient attention to others, and thus regret later
Maybe I should type right-handed again?
Since a freak accident in 2003, I’ve been typing left-handed, but now due to a problem with my computer keyboard, I might have to switch back
Durable nicknames
I revise my earlier theory on what nicknames survive and what don’t. Now I think it has to do more wth the person than the name itself.
Shaking things up
When you think using tools, using a new tool can help you shake up your thinking and get new breakthrough ideas
I’ve started an AI-drafted blog
This one is about Bangalore weather
Interpersonal important and urgent work
At a society and organisation level we’ve come to prioritise urgent work over important work. Which makes it incredibly hard to do well on important but not urgent work
Totes not adorbs
I found India-made tote bags impossible to use, since their handles are way too short. No such issues with foreign totes
Cat-monkey, and LLM work
One prompt takes too long to get answered, and so you give another prompt in another project. Now that takes too long, and soon you’re not sleeping!
Why LLMs mean you need fewer people
There is much greater returns to “intrapersonal vertical integration” now. A lot of the old specialisations and silos will go away since they make LLM-driven programming inefficient
Tukey and Data Exploration
Having learnt it the hard way recently, one thing I'm NOT going to outsource to AI is data exploration. And this is based on an idea from a 1977 book by John Tukey!
The slippery slope of avoiding AI-like writing
People will continue to remove "AI tells" from their writing. AI models will evolve to "write less like AI". New AI quirks appear, which people will distance from. And the cycle goes on...
How LLMs help me think better
They remind me of old (documented) thoughts, and can sometimes develop surprisingly good connections between them
Claude helped me buy coffee
In all my experiments I kept buying suboptimal coffee, so got Claude to build a recommendation system to help me shop better
Abundance and attention
With AI, it is easy to be abundant in your production, but the user's attention is still limited. So the creator's job is to curate, and direct the user on where to pay attention
Open source is a flex
I bring together Thinking Machines (the AI company), Databricks and Taaza Thindi (a South Bangalore restaurant) to talk about how open source can be a massive flex.
Speed of thought and mode of communication
How you choose to write - with pen on paper, or on a computer, or use a scribe, or use voice typing - depends on your speed of thought generation for whatever you're looking to write.
LLMs, Brahma, and one-way doors
Maybe because they are so verbose, only LLMs can read LLM-authored documents or code
Why GLP-1 drugs have a traction problem in India
Indians look down on "taking the easy way out" which is exactly what GLP-1s are seen as. And they are also "artificial, Western innovations", which doesn't help
Human context rot
What I used to call "Bayesian recognition" is, I realise, a special case of "human context rot". And we can use our understanding of LLMs to understand our own thinking better!