Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Phind Model beats GPT-4 at coding, with GPT-3.5 speed and 16k context

phind.com

Post reply on HN
Print view View original

271–280 of 358 posts

Re: Phind Model beats GPT-4 at coding, with GPT-3.5 speed and 16k context

#271

Post by SirMaster »

jiggawatts wrote: post #235
SirMaster wrote: post #58

Earlier quoted context omitted.

It's definitely not impossible at least. Someone is doing it in python here: https://pyatv.dev/ GPT-4 actually sent me here: "Here is an example of a C# library that implements the HAP: CSharp.HomeKit ( https://github.com/brutella/hkhomekit ). You can use this library as a reference or directly use it in your project." Which, to no surprise based on my experiences with LLMs for programming does not exist and doesn't…

In a sense you’re asking it the wrong questions. It’s a bit like asking Google “my PC crashed, how do I fix!?” and then expecting something specific to a rare issue in the first hit. Assuming a C# library even exists for what you’re doing (maybe not!) then still the best use of AI is to troubleshoot specific issues given an almost working piece of code as input . Ask it to explain why something doesn’t work instead o…

Even all of that is on the wrong track. There is nothing that I can see anywhere about controlling an ATV with the homekit accessory protocol.

Re: Phind Model beats GPT-4 at coding, with GPT-3.5 speed and 16k context

#273

Post by SirMaster »

rushingcreek wrote: post #249
SirMaster wrote: post #246

Earlier quoted context omitted.

Unfortunately there’s nothing correct about that answer. There’s no tcp service listening for requests like that on port 7000 on an AppleTV.

It's quoting that from a StackOverflow post: https://stackoverflow.com/questions/11857130/tcpclient-or-ht... .

Yeah, that port 7000 service is AirPlay protocol and they are sending photos and videos to an ATV with it.

But I want to control a unit like send navigation controls like 4 directions, back and select.

The only app I know that can do it is pyatv, (a python app) but I want to do it in my C# app.

It would be nice if an LLM could port pyatv to C# for me as I don't really know python at all.

Re: Phind Model beats GPT-4 at coding, with GPT-3.5 speed and 16k context

#274

Post by jeswin »

iandanforth wrote: post #201
drcode wrote: post #173

I am a heavy user of GPT4, and Phind was surprisingly able to match GPT4 on several initial programming tasks I gave it. Given the large context window of Phind, it will likely be able to outperform GPT4 for some tasks. That is quite an accomplishment, I am impressed

FWIW The default context window of GPT-4 via ChatGPT is about to change to 32k.

Given the number of times it just fails with large prompts on 32k contexts, I'm not sure if they'ready for this. In my experience, if you're consuming 20k+ tokens failure rate is more than 50%.

Re: Phind Model beats GPT-4 at coding, with GPT-3.5 speed and 16k context

#275

Post by antupis »

rushingcreek wrote: post #220
drcode wrote: post #219

Earlier quoted context omitted.

that would put them significantly ahead again, for my use cases

We will eventually increase the Phind Model to 100K tokens -- the RoPE embeddings in Code Llama were designed for this.

100k tokens and good ide support would be great. Copy pasting back and forth with browser and IDE is kinda annoying and you always miss some context. I think model is now good enough but what is kinda missing is good developer experience eg what to load in that context window and how model integrates to IDE. But this is kinda missing with copilot and chatgpt4 as well.

Re: Phind Model beats GPT-4 at coding, with GPT-3.5 speed and 16k context

#276

Post by sagarpatil »

I've been using GPT-4 for coding since it's launched. I did try Phind when you launched it with GPT4 and found it useful but I use VS Code and hate having to switch. I see you do have a VS Code extension now (great job!). I tried the phind model and at first glance it definitely looks better than GPT-4 wrt coding. Do you have any plans to provide this model as an API?

Re: Phind Model beats GPT-4 at coding, with GPT-3.5 speed and 16k context

#277

Post by sagarpatil »

elif wrote: post #257

Hmm I wonder what kind of code quality can be accomplished by looping from phind->gpt->copilot for multiple iterations, asking for criticisms of the code qualities then seeking code which addresses the ai-generated criticisms etc until it knocks out better than I would in a second.

You could cook something like this using Microsoft Autogen. It does allow you to daisy chain the models.

Re: Phind Model beats GPT-4 at coding, with GPT-3.5 speed and 16k context

#278

Post by visarga »

matsemann wrote: post #127
heavyarms wrote: post #85

Earlier quoted context omitted.

I've been thinking along the same lines. The token window IMO should be a conceptual inverted pyramid, where there most recent tokens are retained verbatim but previous iterations are compressed/pooled more and more as the context grows. I'm sure there's some effort/research in this direction. It seems pretty obvious.

But some of the earlier tokens are also the most important ones, right? Like the instructions and rules you want it to follow.

Phrase embeddings could bring a 32x reduction in sequence length because:

> Text Embeddings Reveal (Almost) As Much As Text. ... We find that although a naïve model conditioned on the embedding performs poorly, a multi step method that iteratively corrects and re embeds text is able to recover 92% of 32-token text inputs exactly. We train our model to decode text embeddings from two state of the art embedding models, and also show that our model can recover important personal information (full names) from a dataset of clinical notes.

https://arxiv.org/abs/2310.06816

Re: Phind Model beats GPT-4 at coding, with GPT-3.5 speed and 16k context

#279

Post by letitgo12345 »

dontreact wrote: post #192
rushingcreek wrote: post #73

Earlier quoted context omitted.

Seemed to give plausible results for me: https://www.phind.com/search?cache=lswmiuewv2l33jt337dgrsho

def calculate_relative_advantage(interest_rate, loan_size, market_return): # Your calculation logic here pass Chat gpt actually implements it

Just prompt it to implement the function

Re: Phind Model beats GPT-4 at coding, with GPT-3.5 speed and 16k context

#280

Post by Capricorn2481 »

rushingcreek wrote: post #200
Capricorn2481 wrote: post #190

Earlier quoted context omitted.

Can you share what your long term monetization model is? I'm noticing Phind is free to use right now.

We have a Pro plan where you can get (virtually) unlimited GPT-4 and soon, an even faster Phind model. https://phind.com/plans

Is there something you're doing with GPT-4 that would make me want to use it through you vs just using it myself?
Post reply on HN