sophiabits

8 ms·

1.
by sophiabits 3mo ago
> You can't fire Claude if it fucks up What's the difference between "firing" Claude vs moving to a model from a different provider? The latter seems very analogous to firing an employee for performance and backfillin
2.
by sophiabits 4mo ago
TIL you can get RSS feeds for YouTube channels. Thanks for this!
3.
by sophiabits 4mo ago
> Selfishly, I don't want my screen scraped because it feels like an invasion of my privacy, [...] But zooming out, I don't want to live in a world where humans—employees or otherwise—are exploited for their training data. I wo
4.
by sophiabits 4mo ago
I do not envy the position the npm team are in. They removed the ability to unpublish packages as a response to the left-pad incident[1] because it wasn't desirable for individual developers to break downstream dependencies by pulling
5.
by sophiabits 5mo ago
A browser is OOM more expensive to run than a terminal app, regardless of what you're running inside said browser
6.
by sophiabits 5mo ago
Purchasing power is probably a better metric in a vacuum, but it's hard to analyze accurately For example, the comment you're citing is claiming that because minimum wage has increased only 3x over the same period of time in which
7.
by sophiabits 7mo ago
> the MCP server is automatically launched when the Agent loads that skill The main problem with this approach at the moment is it busts your prompt cache, because LLMs expect all tool definitions to be defined at the beginning of the co
8.
by sophiabits 1y ago
Comments are often the best tool for explaining why a bit of code is formulated how it is, or explaining why a more obvious alternate implementation is a dead end. An example of this: assume you live in a world where the formula for the cir
9.
Agentic checkout in ~100 lines of Python(sophiabits.com)
2 points by sophiabits 1y ago|0 comments
10.
A deep dive into OpenAI's Structured Outputs(sophiabits.com)
1 points by sophiabits 2y ago|0 comments
11.
by sophiabits 2y ago
I’ve especially noticed this with gpt-4o-mini [1], and it’s a big problem. My particular use case involves keeping a running summary of a conversation between a user and the LLM, and 4o-mini has a really bad tendency of inventing details in
12.
by sophiabits 2y ago
I wanted to document a particular genAI antipattern which I've seen a few times now. LLMs are theoretically pretty fungible, because you send English and get English back--but in practice you still need to do some amount of technical d
13.
New LLMs aren't always better(sophiabits.com)
1 points by sophiabits 2y ago|1 comments
14.
by sophiabits 2y ago
Even with this setup in place you need a heightened level of caution relative to a monolith. In a monolith I can refactor function signatures however I desire because the whole service is an atomically deployed unit. Once you have two indep
15.
by sophiabits 2y ago
Not sure what the situation is like now, but we stopped using LangChain last year because the rate of change in the library was huge. Whenever we needed to upgrade for a new feature or bug fix we’d be 20~ versions behind and need to work th
16.
by sophiabits 2y ago
The other possibility (which is common in startups) is that often the “right way” is different depending on the scale of the system you need to design for. In cases like this you end up with technical debt a year down the line, but at the t
17.
by sophiabits 2y ago
Big fan of this. Here in New Zealand we have a slogan “be a tidy kiwi” that encourages people to pick up their litter and be good stewards of for our natural environment Imo the same mentality is good to have in software, and I’ve always ap
18.
by sophiabits 2y ago
This already exists (in a slightly different prompt format); it's the underlying idea behind ReAct: https://react-lm.github.io As you say, I'm skeptical this counts as AGI. Although I admit that I don't have a par
19.
by sophiabits 2y ago
This sounds like an argument against TypeScript in general, no? e.g. If I am parsing a string to a number via Number.parseInt, I don’t need a “: number” annotation because I can just call the variable “myNumber” and use that. Branding a str
20.
by sophiabits 2y ago
Happens a lot with junction tables ime. e.g. At my last job we had three tables: user, stream, user_stream. user_stream is an N:N junction between a user and a stream A user is free to leave and rejoin a stream, and we want to retain old da
21.
by sophiabits 3y ago
Type-prefixed IDs are the way to go. For completeness it's worth noting that the first example using the `string | 'currentNode'` type can be slightly improved in cases where you _do_ want autocomplete for known-good values
22.
by sophiabits 3y ago
There’s a French think tank called “The Shift Project” which actually produced a report estimating CO2 impact of data transfer a few years ago [1]! The numbers are obviously very rough; there’s a LOT of factors to consider and which vary fr
23.
by sophiabits 3y ago
The problem is that you don’t really know if your mock accurately implements the behavior of the database. The only way you could verify the mock behaves correctly would be to run the mock and a real database instance through a set of tests
24.
by sophiabits 3y ago
Yeah, the folks working on aider (AI pair programming) [1] found that these kind of tricks reduced performance for them. I’m pretty confident there will be situations where you can measure a statistically significant performance improvement
25.
by sophiabits 3y ago
Yeah +1 We originally started out building features with LangChain (loading chains from YAML sounded good—it felt like it would be easy to get non-engineers to help with prompt development) but in practice it’s just way too complicated. Nic
26.
by sophiabits 3y ago
Instead of inventing your own DSL for filtering, there are also some well-specified standards. ODATA [1] is probably the main one, and it’s adopted by e.g. Microsoft’s Azure APIs. JSON:API [2] is another, although I’ve never seen it used an
27.
by sophiabits 3y ago
There’s a project called “dprint” which you can use in the interim if Prettier is a bottleneck :) It’s a Rust-based code formatting platform with a Prettier plugin (among others). It’s a lot faster than running Prettier directly [1] The dpr
28.
by sophiabits 3y ago
Adding on to the other comment about CI setups, when I benchmarked removing eslint-plugin-prettier from my work’s ESLint configuration we saw a 37.5% speedup. On our fairly slow CI agents that ended up being pretty significant (~34s per run
29.
by sophiabits 3y ago
This is pretty cool! App router is pretty hard for me to visualize when looking at a file tree (ditto for SvelteKit—I find it hard to grok all the +page.svelte or page.tsx files) Visualizing it all as a tree structure feels like an improvem
30.
by sophiabits 3y ago
Also as a developer I really dislike UUIDs because the hyphens make it impossible to double-click copy. It’s a small gripe, but I need to copy IDs multiple times per day and it adds up. Other ID formats like cuid or KSUID don’t have hyphens

More ›