afro88

24 ms·

361.
by afro88 2y ago
Probably the sender has to be linked to the current code for it to be accepted
362.
by afro88 2y ago
Hard agree. The user being part of the flow is still very much needed. And I have also had a great experience using Vercel's AI SDK on next.js to build an LLM based application
363.
by afro88 2y ago
Semi autonomously is the key. Aider and Plandex are both pretty good.
364.
by afro88 2y ago
OP isn't training though. They're just paying for query. Much cheaper (and insanely faster) than fiverr
365.
by afro88 2y ago
I'm pretty sure they do it client side. The latency on your video preview is non existent.
366.
by afro88 2y ago
Sure, but go to one of the main porn websites and you're shown extremely hardcore content on the front page, in video thumbnails. Often there's something "rough", probaby something about step siblings, and in most cases
367.
by afro88 2y ago
Maybe this post wasn't the right one for your comment, hence the downvotes. But I find it intriguing. Do you mean architecting software to allow LLMs to be able to modify and extend it? Having more of the overall picture in one place (
368.
by afro88 2y ago
That doesn't address it. That tones it all the way down to a "slight element of anger"
369.
by afro88 2y ago
They should still expose the emotion level to the call center operator, just in a different way. Have something on the UI that indicates the level of emotion. It's still an important metric to take into account when resolving the custo
370.
by afro88 2y ago
> Well, yes, but Not all models need to be "super large." Smaller models, specialized in specific tasks, working together - and then reporting to a slightly larger model is the way to go. I want to believe, but I'm still y
371.
by afro88 2y ago
> Over the past few days, we have made other significant improvements to the security of the Spaces infrastructure, including completely removing org tokens (resulting in increased traceability and audit capabilities), implementing key m
372.
by afro88 2y ago
Ahh I see. The time saving part of having it as a chatbot is the way you can easily forward things to it for it to execute a workflow. I thought you meant more like a way of constructing or managing workflows.
373.
by afro88 2y ago
4o can only output 4k tokens. So the training to complete an answer within 4k tokens is probably kicking in and nerfing the quality
374.
by afro88 2y ago
Can you give some examples?
375.
by afro88 2y ago
What's the advantage of using this over PostHog?
376.
by afro88 2y ago
It can work for more complex tests, but you have to give it an initial test that already sets everything up and utilizes mocks correctly. From there it will generate mostly correct tests.
377.
by afro88 2y ago
> POSIWID You need to be honest about what it actually does then. Cherry picking the thing you don't like and ignoring the rest will bring you no closer to true understanding
378.
by afro88 2y ago
Ha nice. I decided to try the same and it got it incorrect. It's so non-deterministic. It landed on 0.17cm. Tried it another time and it got 0.1697cm. When I asked it to check it's work, it got the right answer 0.00589cm
379.
by afro88 2y ago
The same guy found 4o to be much better GPT-4o takes #1 and #2 on the Aider LLM leaderboards https://news.ycombinator.com/item?id=40349655 Subjectively, I've found Aider to be much more useful on 4o. It still makes mis
380.
by afro88 2y ago
Came here to say this. The article assumes that you're already an expert on what you're asking the LLM to help you brainstorm for. A better title would be "LLMs are not suitable for brainstorming in topics you know deeply&quo
381.
by afro88 2y ago
Not sure why you've been downvoted. Needle in a haystack testing exists for a reason
382.
by afro88 2y ago
Yes you're right, it's 4o text but the old audio
383.
by afro88 2y ago
I have access to GPT-4o text and audio, but no video. This is on the iOS app with ChatGPT Plus subscription. Initial connection for audio fails most of the time, but once it's connected it's stable. Sometimes a bit more latency th
384.
by afro88 2y ago
Ah I see now, you took issue with the example specifically, rather than the point. You're right, I see how it's a bad example, and what you've laid out is a very good way of handling dependencies.
385.
by afro88 2y ago
You do need to understand the pros and cons, and wow there are some incredible time and complexity costs to maintaining all dependencies yourself, or trying to firewall all dependencies through interfaces. I wasn't talking about blindl
386.
by afro88 2y ago
> If the technical debt is a problem, 1) we shouldn’t have put it in there I guess it depends on your definition of tech debt, but there are lots of examples of things that aren't tech debt at the time but become tech debt over time
387.
by afro88 2y ago
If self hosted Devv for Teams supports BitBucket, Confluence, JIRA and Azure DevOps, the company I work for (v large enterprise) would be incredibly interested.
388.
by afro88 2y ago
You might be trying to solve problems that they don't really have (but think they have), or you're solving them in the wrong way. For example, is a RAG chatbot the right way to get the information they need from documents, or do t
389.
by afro88 2y ago
Why not measure on diversity of thought then?
390.
by afro88 2y ago
Came here to say this. Thought I missed something! 400B does look set to meet GPT-4, which will be exciting, but it's not finished yet.

More ›