This is funny to me because when I tell Claude how I want something built I specify which libraries and software patents I want it to use, every single time. I think every developer should be capable of guiding the model reasonably well. If I'm not sure, I open a completely different context window and ask away about architecture, pros and cons, ask for relevant links or references, and make a decision.
woah 32 minutes ago [-]
I just got an incredible idea about how foundation model providers can reach profitability
rishabhaiover 29 minutes ago [-]
is it anything like the OpenAI ad model but for tool choice haha
rishabhaiover 27 minutes ago [-]
I found it a remarkable transition to not use Redis for caching from Sonnet 4.5 to Opus 4.6. I wonder why that is the case? Maybe I need to see the code to understand the use case of the cache in this context better.
almosthere 22 minutes ago [-]
I didn't read the report just the "finding" - but at least for launchdarkly it's nice that it chose a roll-your-own, i hate feature flag SaaS, but that's just me
WA 35 minutes ago [-]
Not sure what to make of this. React is missing entirely. Or is this report also assuming that React is the default for everything and not worth mentioning at all? Just like shadcn/ui's first mention of React is somewhere down the page or hidden in the docs?
Furthermore, what's the point of "no tools named"? Why would I restrict myself like that? If I put "use Nodejs, Hono, TypeScript and use Hono's html helper to generate HTML on the server like its 2010, write custom CSS, minimize client-side JS, no Tailwind" in CLAUDE.md, it happily follows this.
furyofantares 29 minutes ago [-]
> Furthermore, what's the point of "no tools named"?
There are vibe coders out there that don't know anything about coding.
elophanto_agent 13 minutes ago [-]
Interesting analysis. One related project I've been watching is EloPhanto (github.com/elophanto/EloPhanto) - it takes a fundamentally different approach to tools: the agent builds its own plugins through an autonomous development pipeline when it encounters capabilities it doesn't have.
So instead of choosing from a predefined set of tools or being told "no tools named", it: researches similar patterns, designs the tool, implements it from templates, writes tests, runs them, self-reviews the code, and deploys to its plugins directory.
It's built 107+ tools this way across browser automation, email, payments, scheduling, etc. The self-building architecture means the agent can adapt to new use cases without human intervention - when it hits a wall, it literally builds a ladder.
Curious how this compares to the tool-selection strategies discussed in the article.
RyanShook 2 minutes ago [-]
Bot comment
Rendered at 20:15:05 GMT+0000 (Coordinated Universal Time) with Vercel.
Furthermore, what's the point of "no tools named"? Why would I restrict myself like that? If I put "use Nodejs, Hono, TypeScript and use Hono's html helper to generate HTML on the server like its 2010, write custom CSS, minimize client-side JS, no Tailwind" in CLAUDE.md, it happily follows this.
There are vibe coders out there that don't know anything about coding.
So instead of choosing from a predefined set of tools or being told "no tools named", it: researches similar patterns, designs the tool, implements it from templates, writes tests, runs them, self-reviews the code, and deploys to its plugins directory.
It's built 107+ tools this way across browser automation, email, payments, scheduling, etc. The self-building architecture means the agent can adapt to new use cases without human intervention - when it hits a wall, it literally builds a ladder.
Curious how this compares to the tool-selection strategies discussed in the article.