Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
redhale
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
10 ms
·
121.
▲
by
redhale
1y ago
Nothing is stopping you from using the API directly, if you prefer to donate more money to Anthropic. I would not personally, as I can't spend thousands per month on an agentic tool. I hope they figure out limits that work. $100 /
122.
▲
by
redhale
1y ago
I didn't have a name for it, but I evolved to this same exact system myself. I use HeyNote [0] for mine. [0] https://heynote.com/
123.
▲
by
redhale
1y ago
Agree. I stopped reading after the blurb below because it tells me this person has not actually even used Copilot or Cursor to a serious degree. This is an AI-written sentence that seems fine, but is actually complete nonsense. > Each to
124.
▲
by
redhale
1y ago
The article has a political focus, but I think anthropomorphism is counter-productive from a technical perspective as well. "Agents" are just prompt completions (or chains of prompt completions) in a loop. "Tool calls" a
125.
▲
by
redhale
1y ago
https://github.com/intellectronica/ruler Library to help with this. Not great that a library is necessary, but useful until this converges to a standard (if it ever does).
126.
▲
by
redhale
1y ago
Agree and disagree. Yes, the bottleneck is going to still be human review. But with only a single Claude Code (or Cursor), I spent a lot of time (a couple minutes at a time) waiting for the agent to run through its loop completing what I as
127.
▲
by
redhale
1y ago
Came here to shout out Crystal. It's the best of these tools I have come across!
128.
▲
by
redhale
1y ago
My second favorite show of all time (second only to Arrested Development). Many shows have memorable/quotable lines, but 30 Rock has more memorable lines per season than many shows do in their entire run. It is rapid fire on another le
129.
▲
by
redhale
1y ago
Fair enough. I guess a better way to put it is: for people who like Cursor's UX, but prefer Claude Code's performance as an agent, the combination of both would be the true killer app. Having to choose between these feels like a t
130.
▲
by
redhale
1y ago
BAML is a language that gives you jinja-like templating. But it also provides structured output parsing and bindings in Python and TypeScript that allow you to execute prompts as typed functions in your code. They also have a VS Code "
131.
▲
by
redhale
1y ago
Kubernetes. AWS. React. All have high learning curves to use effectively, hoardes of footguns, etc. But if you get over that curve, they can be fantastic tools with tons of value. LLM-assisted dev tooling is similar, in my view.
132.
▲
by
redhale
1y ago
+1 for BAML. I find that the "prompts as typed functions" concept really simplifies the mental model here, making LLM apps easier to reason about.
133.
▲
by
redhale
1y ago
Similar: https://www.promptfiddle.com/ From the BAML team, so it uses the BAML syntax (open source).
134.
▲
by
redhale
1y ago
TIL about UUIDv7 -- thanks!
135.
▲
by
redhale
1y ago
Cursor's multi-file tab completion and multi-file diff experience are worth $20 easily IMO. I truly do not understand people's affinity for a CLI interface for coding agents. Scriptability I understand, but surely we could agree t
136.
▲
by
redhale
1y ago
I think you're making the author's point, though. If two users ask the bot the same question and get different answers, is the bot valuable? A dice roll that might be (or is even _probably_) correct is not what I want when going d
137.
▲
by
redhale
1y ago
> Why bother developing chatbots or doing sales, when you will be operating AGI in a few short years? To fund yourself while building AGI? To hedge risk that AGI takes longer? Not saying you're wrong, just saying that even if they d
138.
▲
by
redhale
1y ago
> But whose creative output consists solely of doing things a million people have already done? Is everyone else working on projects built exclusively out of lists of primes and rebalancing binary trees? Yes. Most developers in the corpo
139.
▲
by
redhale
1y ago
In my experience the longer thinking time before acting is due to the model (2.5 Pro). Claude Sonnet 4 is just snappier overall. Same experience in tools like Cursor when comparing these models.
140.
▲
by
redhale
1y ago
Is LiteLLM integration on the roadmap? I get that part of the motivation to build this is to funnel usage to Google models, but it would be nice to be able to go through OpenRouter and/or use local models too. Open source is nice, but
141.
▲
by
redhale
1y ago
Why not both? Claude Code's agent with Cursor's UI harness would be killer, I think. The current CC IDE integration is a poor approximation, imo. The Cursor diff experience is good, as is its tab complete. Claude Code's agent
142.
▲
by
redhale
1y ago
This line by the author, in response to one of the comments, betrays the core of the article imo: > The quality of the code these tools produce is not the problem. So even if an AI could produce code of a quality equal to or surpassing t
143.
▲
by
redhale
1y ago
> I reached out to Tellman before writing this post for some clarifications. Ironic!
144.
▲
by
redhale
1y ago
I also don't understand what your product actually does. What would help me: an example + diagram of a solution before and after the use of Ductape. How does this change things? The more detailed, the better (less marketing language, m
145.
▲
by
redhale
2y ago
100% this. I pay for YT Premium so I have YT Music for free, and I still choose to pay for Spotify because the YT Music app is that bad. Spotify's app is not perfect by any stretch, but comparatively it's amazing. I really miss th
146.
▲
by
redhale
3y ago
"Works on my skull" I typically press it firmly against my chin, and I have observed increased range many times (including testing this trick explicitly with a few different fobs a few years ago out of curiosity). Try pressing mor
147.
▲
by
redhale
3y ago
People on LinkedIn are real. If you use it like a communication tool, it can be very effective at locating opportunities or helping others do the same. I have thousands of loose connections on LinkedIn, many of whom I would/could never
148.
▲
by
redhale
3y ago
LOL You can cheat at Monopoly by just taking more money out of the box. You can cheat at Wordle by googling the answer for that day.
149.
▲
by
redhale
3y ago
You seem to be mistaken: https://9to5mac.com/2023/11/16/apple-rcs-coming-to-iphone/
150.
▲
by
redhale
3y ago
Actual title: "I Made a Link-Sharing Website at the End of 2023"
More ›