parth mistry
@phm
parth mistry
You give the model a workspace and rules:
This feels less like chatting and more like working with a junior analyst who shows their work.
- Inspect the corpus
- Decompose the problem
- Solve sub-parts independently
- Verify logic and assumptions
- Commit only when confidence is high
This feels less like chatting and more like working with a junior analyst who shows their work.
R.I.P. Basic Prompting
remember to check RLM
Humanity has a remarkable ability to ignore time costs.
Evan Armstrong • How To Make Taste Worth $100 Billion
who has better domain knowledge, the vendor or the customer?
🔮 Exponential View #559: coherent agents; goodbye SaaS; orchestrators needed; space phages, immortality & Kimi's …
A a strong case for the death of SAS, which means a reduction in enterprise spend on SAS. Traditional SAS drops because the tone, specificity, and configurations can be defined by the user directly while the user builds software engine types.
Leverage . LLMs are exceptionally good at looping until they meet specific goals and this is where most of the "feel the AGI" magic is to be found. Don't tell it what to do, give it success criteria and watch it go. Get it to write tests first and then pass them. Put it in the loop with a browser MCP. Write the naive algorithm that is very likely... See more
Andrej Karpathy • Tweet
Advanced technology does not move only on sentiment; it moves when prototypes become systems, when pilot projects become assets, and when balance sheets begin to carry the weight of factories, grids, supply chains, and regulated distribution. This shift is measurable. It appears in round sizes, project scopes, unit capacities, offtake terms,... See more
Deep Tech Startups & Venture Capital: An Analysis of 2025
Software stocks are trading at P/E ratios at ten-year lows while their current fundamentals remain strong. That is precisely the signature of a market repricing terminal value, not current earnings. Whether it is overdone depends on whether the next wave of earnings calls shows actual churn or accelerating growth despite the fear.
State of AI: February 2026 newsletter
While I don’t expect diffusion models to supplant autoregressive models, I think there will be more interest in this space, with at least one established lab (Chinese or American) releasing a diffusion-based LLM for mainstream use.
Timothy B. Lee • 17 predictions for AI in 2026
a new kind of llm tech. How does this change the dynamics.
When Generative AI first emerged, SaaS firms’ playbook was to use foundation models from hyperscalers or OpenAI/Anthropic, then train and fine-tune them for specific use cases. The market now believes that, rather than augmenting software, Agentic AI may directly perform tasks for users ( bypassing the need for traditional software ). Since SaaS... See more
Notes on SaaS
The story unfolding here means that rather than incumbents, the native AI company's disruptors are the market leaders.
## Context
I have a recurring [FREQUENCY] task: [BRIEF DESCRIPTION].
Currently I do it manually by [CURRENT PROCESS - 1-2 sentences].
Here’s an example of this task in action:
<example>
Input I provided: [PASTE ACTUAL INPUT]
Output I needed: [PASTE ACTUAL OUTPUT OR DESCRIBE]
</example>
## What I Need
Turn this into a reusable system with:
1.... See more
I have a recurring [FREQUENCY] task: [BRIEF DESCRIPTION].
Currently I do it manually by [CURRENT PROCESS - 1-2 sentences].
Here’s an example of this task in action:
<example>
Input I provided: [PASTE ACTUAL INPUT]
Output I needed: [PASTE ACTUAL OUTPUT OR DESCRIBE]
</example>
## What I Need
Turn this into a reusable system with:
1.... See more
Azeem Azhar • 🔮 Six mental models for working with AI
Remember to use this for your monthly managerial tasks.