nbstme

6 ms·

1.
Why does my harness forget me? Agent engineering(twitter.com)
2 points by nbstme 5mo ago|0 comments
2.
Lessons from Reverse Engineering Excel AI Agents(twitter.com)
2 points by nbstme 7mo ago|0 comments
3.
Every SaaS Is Now an API. Whether They Like It or Not(twitter.com)
1 points by nbstme 7mo ago|0 comments
4.
10 years building vertical software: are we cooked?(twitter.com)
49 points by nbstme 7mo ago|51 comments
5.
The LLM Context Tax: Best Tips for Tax Avoidance(nicolasbustamante.com)
1 points by nbstme 7mo ago|0 comments
6.
Vertical SaaS Is Cooked: The Crumbling Workflow Moat(nicolasbustamante.com)
2 points by nbstme 7mo ago|0 comments
7.
Agentic Search vs. Embedding RAG: An Obituary(twitter.com)
1 points by nbstme 8mo ago|0 comments
8.
Technical lessons from 2 years of building AI agents in financial services(twitter.com)
1 points by nbstme 8mo ago|0 comments
9.
Model Market Fit(nicolasbustamante.com)
76 points by nbstme 8mo ago|13 comments
10.
by nbstme 8mo ago
Product-market fit has a prerequisite that most AI founders ignore. Before the market can pull your product, the model must be capable of doing the job. That's Model-Market Fit. When MMF Unlocks, Markets Explode (legal, coding...)
11.
by nbstme 11mo ago
The best AI code is the code you delete. Models are eating abstractions faster than teams can adapt. Each model introduces a new paradigm shift. If you miss a paradigm shift, you’re dead.
12.
by nbstme 1y ago
After 30 years of clicking, scrolling, and optimizing pixels, websites are becoming obsolete. LLM agents will read and act for us, ending search engines, blue links, and traditional websites.
13.
by nbstme 1y ago
Exactly. Just a markdown file per user. Anthropic recommends that.
14.
by nbstme 1y ago
HAHAHA. Ok let's call it "transformation." As i wrote "The next decade of AI search will belong to systems that read and reason end-to-end. Retrieval isn’t dead—it’s just been demoted."
15.
by nbstme 1y ago
Thanks! Sorry if the flow was off
16.
by nbstme 1y ago
haha so true!
17.
by nbstme 1y ago
Why call it an ad? It’s not even on the company site. I only mentioned my company upfront so people get context (why we had to build a complex RAG pipeline, what kinds of documents we’re working with, and why the examples come from real pro
18.
by nbstme 1y ago
But don’t you think LLM pricing is heading toward zero? It seems to halve every six months. And on privacy, you can hope model providers won’t train on your data, (but there’s no guarantee)
19.
by nbstme 1y ago
Why does grep in a loop fall apart? It’s expensive, sure, but LLM costs are trending toward zero. With Sonnet 4.5, we’ve seen models get better at parallelization and memory management (compacting conversations and highlighting findings).
20.
by nbstme 1y ago
LLMs > rerankers. Yes! I don't like rerankers. They are slow, the context window is small (4096 tokens), it's expensive... It's better when the LLM reads the whole file versus some top_chunks.
21.
by nbstme 1y ago
Appreciate the feedback. I’m not saying grep replaces RAG. The shift is that bigger context windows let LLMs just read whole files, so you don’t need the whole chunk/embed pipeline anymore. Grep is just a quick way to filter down candi
22.
by nbstme 1y ago
Yes my point is that the entire RAG pipeline like ingest, chunk, embed, search with Elastic, rerank is in decline. Grep is far simpler. It’s trivial.
23.
by nbstme 1y ago
Exactly. AGI implies minimal tooling and very primitive tools.
24.
by nbstme 1y ago
hum because Claude Code pioneered the 'grep/glob/read' paradigm, so I felt the need to explain that what works well for coding files can also be applied to more complex documents.
25.
by nbstme 1y ago
100%. The speed of change is wild. With each new model, we end up deleting thousands of lines of code (old scaffolding we built to patch the models’ failures.)
26.
by nbstme 1y ago
Agree. It's a context/memory issue. Soon LLMs will have a 10M context window and they won't need to search. Most codebases are less than 10M tokens.
27.
by nbstme 1y ago
Why? What if LLMs could parallelize much of their reading and then summarize the findings into a markdown file, eliminating the need for complicated search?
28.
by nbstme 1y ago
The main use case I referenced is SEC filings search, which is quite different from code. Filings are much longer, less structured, and more complex, with tables and footnotes.
29.
by nbstme 1y ago
Haha! LLMs themselves are pure edge cases because they are non-deterministic. But if you add a 7-step pipeline on top of that, it's edge cases on top of edge cases.
30.
by nbstme 1y ago
It's mind blowing. It's so simple, elegant and... effective! Grep+glob and a lot of iterations is all we need.

More ›