Day 7 - Agents and agentic work flow - but once you embrace it life becomes a series of orchestration (builder life, not real life).

Day 6 - Context burden. Models, the main have increasing memory about you. So makes assumptions and lives with the same bias you already have. once the model suggested that I don't invest in high risk funds as I already ran a startup.

Day 6 - Chain of though prompting, complex instructions, "this is very important - do not make mistake " kind of promopts actually back fire now. Newer models have more reasoning. They are not just pattern completer

Day 5 - Context window and context decay are something developers need to worry about. In medicine it’s unlikely you will hit it, but still it’s better to be aware and open new chats when needed

Day 4 - models are general-purpose…… To make one behave like a specific product, you constrain it. Every rule is a tradeoff: more rules = more predictable but more brittle and more tokens. This is the actual craft of building AI products.. so we go back to the cost benefit play .

Day 3 - Technical jargons - prompt, toke, context window, context decay - AI use cases are full of these keywords. It’s very important that we deeply understand them. I used to dismiss them with superficial know - how. As someone who used to hate jargons and buzzwords I am learning to like them again (and understand them)

Day 2- When code and design become easy with AI - then your tastes become important. Same logic for everything you do with AI. Example - a slide deck. All decks will look clean and manicured. But your taste will stand out. In that way, I consider myself as the CTO (Chief Taste officer) for my new startup

90 day project. Day 1 - Think about this - ChatGPT and Claude will try to impress you. The model want to impress as they your business. They are more likely to over explain, give more solutions than needed, and generally complicate to get your attention. Be mind full of this and prompt against it.

awesome!

what an achievement. 275 x 2 lives and families. I have celebrated for an app button launch 😂 . Humbling. And very cool! Congratulations

No... it's a point of view of @autopsy_surgeon . I just countered it....... as times have changed -

Kids use all shortcuts. Marrow/prep/btr/short videos /osmosis/chatGPT/Gemini (AI will be big)/uptodate. Dn't bother reading journals - jump into @openevidence & read only if needed. Big books with small fonts are a pain and old. Be efficient and if you have time left - fight against the increase of MBBS seats in India. That will give you a better life. No fat books with small fonts.

7/n

7/n Be resouceful. Have the connections, build the connections. Save the bookmarks, write the notes. Read the books. Have better prompts. Ask for recommendations and then save it. This compounds over your lifetime

6/n Be couragous to see beyond - suspect. Just because everyone does it - is not gonna make anything true or ideal. Remember. Crowd always follow. Crowd never leads. Infact be suspicious when something is done is someway, because everybody does it. So have the guts to see beneatt

5/n Self negotiate - Question everything, your own assumptions. Be the first to change if needed. Don't worry about changing your opinion. Strong opinions loosely held.

4/n Persist. Show up everyday. You will be in top 5 precentile by showing up everyday. Just be the person who won't stop it, because it's boring or repitive or not showing any results in short time.

3/n Bias toward action - Get things done. Do it even if it is not in a perfect state. Be the momentum builder, be the first mover. Know as the person who gets things done. This starts by simply doing things witht weight in cost benefit or thinking what you will get out of it

2/n Be responsible for the OUTCOME. Be the guy who takes the fall. Be the person who is always where the buck stops. It's an internal self culture. Once you are that person - people will trust you with life.

1/n Focus on solution. Don't just crib about the problem. There is always a way. Even if there is no solution in short term, keep fantasising how you will solve in future. Find smaller solutions always. And implement