Leynos

15 ms·

211.
by Leynos 1y ago
Get Coderabbit or Sourcery to do the code review for you. I tend to do a fine tune on the reviews they produce (I use both along with CodeScene), but I suspect you'll probably luck out in the long term if you were to just YOLO the revi
212.
by Leynos 1y ago
Tiny selection of web sites, restricted to the US, absolutely no marketing.
213.
by Leynos 1y ago
Is this companies collecting data for model training, or is it agentic tools operating on behalf of users?
214.
by Leynos 1y ago
Set your temperature to zero
215.
by Leynos 1y ago
I prefer to talk through a code review with the candidate. I like to get a feel for how they approach another developer's code, the type of questions they ask, and whether or not they have a sense of situational needs. It's worked
216.
by Leynos 1y ago
I find that if I generate a Deepresearch paper in Gemini first, then pass that Deepresearch paper to NotebookLM, I get really good results if I don't have the sources to hand first.
217.
by Leynos 1y ago
It's funny seeing this as the top story. I'm in the middle of putting together a realtime strategy game using Differential Datalog[1] and Rust, with DDL managing the game's logic. Mostly as an excuse to expose myself to new i
218.
by Leynos 1y ago
I've been using it for problems where the model struggles. Even if it doesn't find an answer, the extra opportunity for it to explore the problem space usually helps in getting me closer to the right answer.
219.
by Leynos 1y ago
This will be the next benchmark suite: how long does it take your model to interactively retrieve an API key.
220.
by Leynos 1y ago
Would you skip these guardrails when working with a "Junior to Intermediate" developer?
221.
by Leynos 1y ago
Seemingly so going by the sibling comment to yours.
222.
by Leynos 1y ago
Synthetic training data presumably.
223.
by Leynos 1y ago
Did you use search grounding? O3 or o4-mini-high with search grounding (which will usually come on by default with questions like this) are usually the best option.
224.
by Leynos 1y ago
Do you give the agent a style guide? Do you perform (or have another agent perform) code reviews on the agent's code? Do you discuss architecture and approach with the agent beforehand and compile that discussion into a design and deve
225.
by Leynos 1y ago
Something I like to do is get Gemini Deepresearch to write a single file manual for any uncommon dependency and include that in my docs/ directory. Helps a bunch. I also ask it to write specialized guides on narrow topics (e.g., testin
226.
by Leynos 1y ago
Play the game online: https://warrior-rockk.itch.io/out-of-cash-sin-blanca
227.
by Leynos 1y ago
0.002 tokens per second
228.
Ty: Astral's New Type Checker (Formerly Red-Knot) – Talk Python Live Stream [video](youtube.com)
1 points by Leynos 1y ago|1 comments
229.
by Leynos 1y ago
Interesting. It's illegal to drive without insurance in the UK.
230.
by Leynos 2y ago
Why are you letting the AI construct poorly structured code? You should be discussing an architectural plan with it first and only signing off on the code design when you are comfortable with it.
231.
by Leynos 2y ago
And which parts of that governance can be kept secret should be subject to continuous review. Openness must be carefully guarded.
232.
by Leynos 2y ago
A functioning justice system is an important part of a functioning democracy.
233.
Quasar Alpha(openrouter.ai)
1 points by Leynos 2y ago|0 comments
234.
by Leynos 2y ago
I use these in Roo Code in VSCode with Claude Sonnet 3.7.
235.
by Leynos 2y ago
If you want an LLM to use a tool, you just need to implement a parser in your LLM client that extracts the tool call from the LLM's response, then give the LLM a syntax it can use to make the tool call. For example, in Roo Code: ``` TO
236.
by Leynos 2y ago
That article makes very little sense. The protocol itself sits on top of JSON-RPC, and the specifications are there for anyone to implement. There's nothing specific to Claude about it. There are various MCP client and server implement
237.
by Leynos 2y ago
I wrote MCP servers for git, pytest, and ruff, and those have been fantastic.
238.
by Leynos 2y ago
Just consider the free version a demo.
239.
by Leynos 2y ago
Yes, huge monoliths
240.
by Leynos 2y ago
Sadly, qwq failed: > If I was to tell you that the new sequel, "The Fast and The Furious Integer Overflow Exception" was out next week, what would you infer from that? > I'm sorry, but I can't assist with that. Out

More ›