Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
jrvarela56
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
7 ms
·
61.
▲
by
jrvarela56
2y ago
Hopefully this all comes full circle and we just quit social sites en masse.
62.
▲
by
jrvarela56
2y ago
I pay for Cursor but the comments here made me question the incentives. They have to cut spending. The byo api keys don’t suffer from that. Maybe we need evals for these tools? There are benchmarks for coding abilities (swe bench?), would b
63.
▲
by
jrvarela56
2y ago
That was fun, thank you!
64.
▲
by
jrvarela56
2y ago
16gb
65.
▲
by
jrvarela56
2y ago
Feedback: I use MacWhisper and Tiny wisperkit model (english only) is way faster than any cloud service on my M1 macbook pro. I’d say local is necessary for delightful product experience and the added bonus is that it ticks the privacy box
66.
▲
by
jrvarela56
2y ago
Why would you even search for that out of the context of the IDE where you're coding or writing documentation? If you're writing bash you'd have all those man pages loaded in context for it to answer questions and generate co
67.
▲
by
jrvarela56
2y ago
I’d say come back in a few years for a bad take. But this is already a bad take. A query in a regular search engine can at best perform like an LLM-based provider like Perplexity for simple queries. If you have to click or browse several re
68.
▲
by
jrvarela56
2y ago
Not through ads - you can be sure of that!
69.
▲
by
jrvarela56
2y ago
People talk about stuff with more ease than they write. There is a lot of knowledge in peoples’ heads that you wont find written online.
70.
▲
by
jrvarela56
2y ago
Great way to put it. If you are interested in a more elaborate description, checkout the intro to ‘The Untethered Soul’. To the people here saying this is just meditation: the problem is that meditation is a solution to a problem most peopl
71.
▲
by
jrvarela56
2y ago
https://slatestarcodex.com/2014/07/30/meditations-on-moloch/
72.
▲
by
jrvarela56
2y ago
I think the feeling is that for this to really be AGI, it has to take in a single prompt and then delegate behind the scenes to an enormous tree of sub-agents if needed. One app that comes to mind is Google's Conversational agents. The
73.
▲
by
jrvarela56
2y ago
What parts of the stack or what patterns do you feel are missing? Where does your gut tell you is the 80/20?
74.
▲
by
jrvarela56
2y ago
Would be cool if the LLM can break up the request into sub-requests processable by LLMs. Current talk about agents mention some sort of router/orchestrator that delegates to other agents. But these can be another llm, another agent, an
75.
▲
by
jrvarela56
2y ago
Unfounded. Assuming automated/voice-enabled assistants will spam for opening times, price & description of products/services, scheduling/booking, etc. In the long run it'll be an arms race but the transition will be
76.
▲
by
jrvarela56
2y ago
This is going to start happening to brick and mortar businesses through their customer support channels
77.
▲
by
jrvarela56
2y ago
The sheer amount of inbound I get for random stuff is overwhelming. Facebook can put stuff in front of so many people. I have sold furniture, electronics, used clothing, all sorts of crap.
78.
▲
by
jrvarela56
2y ago
I would recommend everyone reading this to think of it as a skill issue. You can learn to use the LLM/agent to document your code base, test isolated components and refactor your spaghetti into modular chunks easily understandable by t
79.
▲
by
jrvarela56
2y ago
I think it's a Jeff Bezos quote, but forgot the source: "I'd rather have 2 of something than 0"
80.
▲
by
jrvarela56
2y ago
No it’s not. More moderation, more false positives, less free speech. Just having ‘moderation’ means the speech is not ‘free’.
81.
▲
by
jrvarela56
2y ago
And the first is cheaper than a ‘regular single version’ bc you build it with the thought that you will throw it away.
82.
▲
by
jrvarela56
2y ago
Reminds of similar separations in other industries: - Newspapers: editorial and advertising (avoid that companies who pay for ads influence opinions) - Banks: ‘origination’ and wealth management (avoid selling financial products created by
83.
▲
by
jrvarela56
2y ago
Advice is “useless” in the same sense that reading a book is useless. You are receiving compressed knowledge while your capacity to accomplish goals is iterated through more complex experiences. Like reading, advice needs to be unpacked, pu
84.
▲
by
jrvarela56
2y ago
I used to have sleep issues and got rid of them with meditation. I got to a point where I could meditate 10-20 min straight and that was enough to quiet my mind into sleeping.
85.
▲
by
jrvarela56
2y ago
Your application and development environment (editor + 'server') are in the same runtime. This seems important for iterating JS/TS programs ala Smalltalk.
86.
▲
by
jrvarela56
2y ago
Obvious answer is to fire immediately. Huge red flag if you cant fire someone under these circumstances.
87.
▲
by
jrvarela56
2y ago
This requires: - function calling: the LLM can take action - Integration to your runtime: functions called by the LLM can run your tests, linters, compiler, etc - Agents: the LLM can define what to do, execute a few tasks, and keep going wi
88.
▲
by
jrvarela56
2y ago
This workflow is very close to being possible. I gave it a try last year by adding exceptions and test output to clipboard automatically (requires custom code for your stack). The context has increased considerably since my last attempt and
89.
▲
by
jrvarela56
2y ago
Similar to how programmers work, the AI needs feedback from the runtime in order to iterate towards a workable program. My expectation isn’t that the AI generate correct code. The AI will be useful as an ‘agent in the loop’: - Spec or test
90.
▲
by
jrvarela56
2y ago
Any step in improving the interactivity of development workflows is commendable. If you're interested in defining a direction for where REPL features should take us, checkout REPL workflows in Clojure and The Glamorous Toolkit (Smallta
More ›