Ask HN: How productive are you with ChatGPT 4 without history? | Hacker News Reader

Ask HN: How productive are you with ChatGPT 4 without history?

3 pointsbraingenious6 comments

I’ve been trying to get gpt4 to help with a basic python script to clean up wine messy html, and due to it erasing the history I’m having to paste the entire script with every few prompts so it doesn’t wander off and write something useless.

However, if I paste in too much it locks out the text box and tells me to refresh.

This resets the history.

For a productivity tool it’s kind of funny having to teach myself how to work around complete feature breakage and service degradation for it. Goldldfish memory more might actually be wasting my time!

View on HN

For me, I expect LLMs to become significantly more useful when enabled with multi-session continuity and access to a persistent store of data personally relevant to me. Of course, these are also things which tend to increase the security, privacy and abuse surfaces of LLMs.

What I hope someone comes up with is an open standard which lets me choose when to allow an LLM private, temporary 'merge access' to my personal data (eg emails, documents, calendar, browsing history, online posts across platforms and accounts, chats, etc). If no vendor offers something like this then I'll probably host my own private LLM based on whatever the best available open source project is.

I agree! I am excited for the future of all this.

Today I managed to get Alpaca 13B running locally on my laptop and I’m really impressed. I look forward to being about to run 65B with my next hardware upgrade, though I suspect something even better will be available by the time I get around to actually buying anything.

Running a decent LLM on a laptop is super impressive though I expect I'd run a personal LLM on a server for 24/7 availability, backup, etc. It'll be interesting to see if people like us HNers end up running our own full LLMs on open source for privacy, security and control reasons or if an acceptable solution is worked out which allows keeping our personal data blobs separate but able to be temporarily securely accessed.

It's pretty much fine, it is still so powerful that I can recreate what I want trivially. If anything this goes to show that you should independently persist any prompts that are important to you!

I certainly will do so going forward! I will be creating an entire local backup scheme of everything to deal with both acute and anterograde amnesia.

Since we must plan for complete loss of historical data and the inability to save data, what’s other errors in core capability are prudent to proactively plan around?

For example, if it’s liable to fail to store either my prompt or its response, maybe it could inaccurately store my prompt or give inaccurate output. How do you plan around past/current hallucinations?

First I think you really want to get parity between the ChatGPT app and some kind of open source client that just interacts with the API. Once you have achieved that, then people could come up with alternative storage mechanisms that work more effectively. I know this is kind of handwavey, but the persistence issues that OpenAI have shown with ChatGPT are really problematic. I've just been recording prompts that are important to me, or prompt templates, but even losing my entire prompt history multiple times (still gone as of today) - I don't feel like it particularly slows me down. I usually use it for what I need, use the output for whatever purpose I need it for, and then go off to do something else.

Reply on news.ycombinator.com