HNHacker News
TopNewBestAskShowJobs

saliagato

380 karma · joined May 31, 2022

submissionscomments
saliagato··on TradingGPT: LLM multi-agent framework for enhanced financial trading performance
Wow. This is the future. I am going to launch a product based on this
saliagato··on What are we calling what used to be called FAANG these days?
That’s exactly what he said
saliagato··on LLMs, RAG, and the missing storage layer for AI
Azure Cognitive Search takes care of all of this combining semantic search with other layers of traditional search methods
saliagato··on Claude Pro
I still find ChatGPT (4) far superior to Claude.
saliagato··on Show HN: Turn Multiple Documents into Comparative Tables with GPT
Very good keep going. I am going to subscribe
saliagato··on Show HN: Turn Multiple Documents into Comparative Tables with GPT
I wouldn’t trust a fyi tld
saliagato··on ChatGPT Enterprise
everything but 32k version and 2x speed is the same as the consumer platform
saliagato··on ChatGPT Enterprise
I thought that Microsoft was busy with enterprises yet OpenAI announces a product for enterprises. I have a feeling that the two do not get along
saliagato··on You probably don’t need to fine-tune an LLM
Not OP but, I finally got access to GPT-4 32k and I thought that for a thing as simple as a summary it would be better than the smaller context-window model. But then I realized that now I have (practically) infinite context in input but what the LLM can output is always between 1-1.5k tokens. That is, I think, because of the samples in its training, not because it can't output a more lengthy output. In short, I think that multiple steps are better, for example in summarization.
saliagato··on Show HN: OpenPilot, an open-source AI coding assistant for VS Code
Cool keep going
saliagato··on Azure ChatGPT: Private and secure ChatGPT for internal enterprise use
How?
saliagato··on Azure ChatGPT: Private and secure ChatGPT for internal enterprise use
This is exactly the same
saliagato··on Azure ChatGPT: Private and secure ChatGPT for internal enterprise use
Yes exactly
saliagato··on MetaGPT: Meta Programming for Multi-Agent Collaborative Framework
I ageee. Other examples can be reproduced with a straight input output prompt
saliagato··on Ask HN: What makes your AI special that can't be replicated by competitors?
Cool. Do you mind me asking what do you mean by “trained”? Fine-tuned on your data?
saliagato··on Llama 32K Context Released by Together AI
You are correct. The paper is called "Lost in the middle" [1] and it is probably one of the worst drawbacks of this technology. It makes a lot of use cases biased (think of law).

[0] https://arxiv.org/pdf/2307.03172.pdf

saliagato··on GPT-4 API General Availability
Probably. They will have different prices to finetune too.
saliagato··on GPT-4 API General Availability
Unfortunately their decisions are driven by model usage: gpt-3.5-turbo is the most used one (probably due to the low price and similar result)
saliagato··on GPT-4 API General Availability
wait a couple of hours
saliagato··on GPT-4 API General Availability
That's what I always thought. Someday they will come up with a new embedding model, right?
saliagato··on Azure OpenAI Service on Your Data in Public Preview
No
saliagato··on Function calling and other API updates
pretty much the same. slightly worse at following instructions
saliagato··on Entrepreneurs who regret starting businesses
“ I have definitely lost a few years of life expectancy “

Why? Genuinely curious. Stress can lower life expectancy?

saliagato··on Italian privacy regulator bans ChatGPT
How is this ban vague? Have you read the full announcement?[1]

Here are the reasons:

DETECTED, from a check carried out in this regard, that no information is provided to users, nor to interested parties whose data has been collected by OpenAI, L.L.C. and processed through the ChatGPT service;

NOTING the absence of an appropriate legal basis in relation to the collection of personal data and their processing for the purpose of training the algorithms underlying the functioning of ChatGPT;

NOTING that the processing of personal data of the interested parties is inaccurate as the information provided by ChatGPT does not always correspond to the real data;

DETECTED, moreover, the absence of any verification of the age of users in relation to the ChatGPT service which, according to the terms published by OpenAI L.L.C., is reserved for individuals who are at least 13 years old;

CONSIDERING that the absence of filters for minors under the age of 13 exposes them to absolutely unsuitable responses with respect to their degree of development and self-awareness;

CONSIDERING therefore that in the situation outlined above, the processing of personal data of users, including minors, and of interested parties whose data is used by the service is in violation of articles 5, 6, 8, 13 and 25 of the Regulation;

RECOGNIZING, therefore, the need to have, pursuant to art. 58, par. 2, lit. f), of the Regulation - as a matter of urgency and pending the completion of the necessary investigation with respect to what has emerged so far against OpenAI L.L.C., a US company that develops and manages ChatGPT, the extent of the temporary limitation of the treatment;

CONSIDERING that, in the absence of any mechanism for verifying the age of the users, as well as, in any case, of the complex of violations detected, said temporary limitation must extend to all personal data of the interested parties established in the Italian territory;

CONSIDERED it necessary to order the aforementioned limitation with immediate effect from the date of receipt of this provision, reserving any other determination to the outcome of the definition of the preliminary investigation started on the case;

RECALLING that, in the event of non-compliance with the measure established by the Guarantor, the criminal sanction pursuant to art. 170 of the Code and the administrative sanctions provided for by art. 83, par. 5, letter. e), of the Regulation;

CONSIDERING, on the basis of the foregoing, that the prerequisites for the application of art. 5, paragraph 8, of Regulation no. 1/2000 on the organization and functioning of the Guarantor's office, which provides that «In cases of particular urgency and in which the Guarantor cannot be convened in good time, the president can adopt the measures pertaining to the body , which cease to have effect from the moment of their adoption if they are not ratified by the Guarantor in the first useful meeting, to be convened no later than the thirtieth day";

[1]https://www.garanteprivacy.it/web/guest/home/docweb/-/docweb...

saliagato··on Italian privacy regulator bans ChatGPT
I’m Italian. Google Translate is better here
saliagato··on ChatGPT Plugins
Information retrieval to the prompt
saliagato··on ChatGPT Plugins
Well that’s a win-win situation
saliagato··on Prompt Engineering: Steer a large pretrained language model to do what you want
wow
saliagato··on ChatGPT has trouble giving an answer before explaining its reasoning
This is actually the key. David Shapiro explored this possibility and created a concept of Natural Language Cognitive Architecture [0]

[0] https://github.com/daveshap/NaturalLanguageCognitiveArchitec...

saliagato··on Google Should Fire Sundar Pichai
can confirm
← PreviousPage 2 of 3Next →