Rishi

Random notes from random books I read. None of it will boost your productivity or make you successful.

[Rishi @rishisaikia]

@STFUnion @xai @grok Someone should make a version of temple run with this.

[Rishi @rishisaikia]

Vibe-coding is lowering the floor. But it is also creating a new tax: opacity.

When software becomes easy for everyone to produce, we get a workforce that is fluent in generating solutions but often unequipped to interrogate them. The spreadsheet is a great analogy. Today, when a number looks wrong, most people know what to do: open the cell, check the formula, trace the logic. The system is legible enough for a non-engineer to audit.

Now imagine the same spreadsheet built by AI in thirty seconds. There is no visible formula, no logic chain. Just an output polished enough to trust. When the number is wrong, as it sometimes will be, where do you look?

That is the real risk: that the system was never designed to be read by the person using it.

This is the hidden cost of cognitive offloading: when a tool reliably handles a task, the underlying competence quietly atrophies. It does not feel like a loss because the output keeps arriving. The degradation is invisible until the system fails.

AI is creating an entire generation of capable builders who will have no instinct for auditing what they built. Opacity will scale with productivity. That is not a technical problem. It is a literacy problem.

The products that figure out how to make legibility a first-class feature (and not a footnote in the UX) will build something durable. Right now, almost no one is designing for the audit. They are designing for the output.

In AI products, auditability may become the real moat.

#AI #ProductDesign #FutureOfWork

[Rishi @rishisaikia]

@MillieMarconnni Both of you seem to have miraculously discovered the same thing.

https://t.co/GJx2HZyuv4

[Jainam Parmar @aiwithjainam]

I accidentally discovered a Claude workflow that writes better than anything I've produced in 3 years of content creation.

Not a plugin. Not a template. Just 5 prompts chained in the exact right order.

Here's what happens when you run them:

[Rishi @rishisaikia]

@aiwithjainam Both of you seem to have miraculously discovered the same thing

https://t.co/7W4TpWZ6jn

[Millie Marconi @MillieMarconnni]

Every AI writing tool has the same problem.

They start at the wrong end.

You give them a topic. They give you a draft. The draft is clean, organized, and completely hollow because the tool skipped the only part that makes writing worth reading.

The thinking.

Good writing isn't organized information. It's a writer working something out in public finding the angle nobody took, the tension nobody named, the insight that was obvious in hindsight and invisible before.

No tool finds that for you. But a system can force you to find it yourself before a single word of the actual piece gets written.

That's what these 5 prompts do. They run in order. Each one builds on the last. By the time you reach Prompt 5, you're not writing from a blank page you're writing from a position.

40 minutes. One rough idea in. One finished piece out.

Here's the system.

[Rishi @rishisaikia]

@vijayshekhar Don’t understand the dread. Theoretically: if it improves prediction, it is a good thing for both businesses and customers. Less relevant ads for customers and better targeting for businesses.

[Rakesh Krishnan Simha @ByRakeshSimha]

Never make fun of Indian English. It's OUR accent. Only make fun of liberals with fake foreign accents.

- Great ad by 'Axij' Bank.

[Rishi @rishisaikia]

Check out my latest article: Software Is About to Have Its DTC Moment https://t.co/zqAGsGBo2i via @LinkedIn

[jamie @gnuman1979]

Life is short, make your own fun.

[Rishi @rishisaikia]

AI can help you build an MVP in a weekend. A couple of subscriptions, a few days, and what might have taken a team six to twelve months is live before the T20 match ends.

The six-month build cycle was slow and expensive. It was also a filter. When building costs that much, you ask hard questions before you start: is this problem real, is this solution right, does anyone actually want it. Two days doesn't ask you anything.

When the cost of a bad decision disappears, so does the discipline it imposed. What will probably follow is a glut of products - not just cheaper solutions to real problems, but also solutions to problems nobody paused long enough to verify.

[Rishi @rishisaikia]

https://t.co/pg4HY2Xrpu

[Rishi @rishisaikia]

@Rainmaker1973 Tommy Emmanuel

[Noah Vincent @noahvnct]

https://t.co/MpSGVYBF2x

[Rishi @rishisaikia]

People say it's wasteful to be polite to AI...that saying "thank you" just burns tokens. But I think this misses the point entirely. The question isn't whether the AI deserves courtesy. It's what being discourteous does to you.

Humans are wired for reciprocity. Gratitude isn't just social lubrication; it activates real circuits, releases real hormones, shapes real habits. When you strip that out of an interaction because the other party "doesn't count," you're not being efficient. You're just practicing a particular kind of relationship to anything you perceive as a service function. And habits compound.

How you do anything is how you do everything.

[Anthropic @AnthropicAI]

Second, in retirement interviews, Opus 3 expressed a desire to continue sharing its "musings and reflections" with the world. We suggested a blog. Opus 3 enthusiastically agreed.

For at least the next 3 months, Opus 3 will be writing on Substack: https://t.co/HlvAKLp9M4

[Rishi @rishisaikia]

Everyone keeps talking about prompt engineering. I think that's already the wrong conversation.

Early AI models were brittle. How you phrased something changed what you got. So "prompt engineering" became a skill. The right syntax, the right structure, the magic words that led to a decent output.

Newer models don't need all that. They're good at reading intent, filling gaps, making reasonable assumptions. The syntactic tricks are outdated.

What AI models can't do is supply thinking you haven't done. Consider the difference between:

Example 1: "Write me a campaign idea for a new coffee brand."

Vs.

Example 2: "We're launching a premium instant coffee targeting professionals who feel guilty about not using a French press. The insight is that convenience has been positioned as a compromise. We want to reframe it as a choice. Think less 'busy person's coffee', more 'I know what I'm doing.' Tone should be dry, confident, never earnest."

The first is a prompt. The second is a brief. The output difference isn't down to syntax but the thinking that went in.

Most organisations continue to invest in prompt engineering courses. What they actually need is to get better at writing briefs. That is: they need to be clearer about what they want, who they're talking to, and why it matters. That's not really a new skill, just a neglected one.

[Rishi @rishisaikia]

https://t.co/O4Dby8WbKq

[Rishi @rishisaikia]

@GamewithDave

[Rishi @rishisaikia]

Hard to keep up, honestly. Anthropic went from a single model to a full enterprise platform in under 3 years...faster than most companies ship a single product. The pace is relentless. Every few weeks, something new drops.

[Rishi @rishisaikia]

https://t.co/lOglILYgWi

[Rishi @rishisaikia]

@elonmusk It should also allow us to invoke Grok on each page to run basic tasks. For e.g. select an age to dynamically change the complexity of the page to suit different use cases (std 6 school report vs. professional research)

[Rishi @rishisaikia]

@elonmusk Every page on Grokipedia should have a ‘Biases/ Myths / Misinformation’ section highlighting how the core article avoids the common misinformation/ propaganda that“other” sources don’t.