Qredence.ai

Adaptative Intelligence Systems / @zacharybydesign @agenticfleet #NVIDIAInception @msft4startups

[Zachary BENSALEM @ZacharyByDesign]

Introducing a series on @AgenticFleet Fleet RLM, the foundation of the Agentic System leveraged through RLMs craft at its core @DSPyOSS and @daytonaio .

For this eval, 100 runs, LongCot Mini, No Pre-Training, No Optimization, and using @deepseek_ai V4 Flash vanilla , and with Fleet RLM

[Zachary BENSALEM @ZacharyByDesign]

https://t.co/fClRXoYqSU

[Zachary BENSALEM @ZacharyByDesign]

Are you using @cursor, the AI code editor? To further boost your productivity, check out the meta-prompts I created to make Cursor ultra-efficient!

Visit the official page: https://t.co/cj8trymeOs

For a quick demo of one of the three rules:

#cursor #IDE #rules

[Zachary BENSALEM @ZacharyByDesign]

We've released GraphFleet Alpha 0.2.0! Enjoy a smoother, more interactive data indexing workflow. #GraphFleet #RAG

✨ New Jupyter Notebook integration! Upload PDFs, set output directories, and run the GraphFleet indexing pipeline seamlessly.

https://t.co/nNSWiK0FEu

[Zachary BENSALEM @ZacharyByDesign]

What if the AI Chat Protocol from @Azure could integrate with @figma and sync with a repo? 👀

No more battling with code if you're not a developer. Maintain structure as needed with your own interfaces.

#FleetUI from @QredenceOS is coming very soon!

#buildinpublic

[Zachary BENSALEM @ZacharyByDesign]

So far the @microsoftfrance #Build conference is awesome with lots of things whereas in preview or available, that will foundamentally help the pace toward @QredenceOS open-sourced line of solution (cloud & self hostable)
Coming Soon.

[Shaokun Zhang @ShaokunZhang1]

Excited to announce the acceptance of our paper titled "Training Language Model Agents without Modifying Language Models" (title change to “Offline Training of Language Model Agents with Functions as Learnable Weights” in the revised version.) at #icml2024 @icmlconf

1/N

[Zachary BENSALEM @ZacharyByDesign]

Experience the mind-blowing power of @GroqInc's Llama 370b in a @weweb_io app! Say goodbye to streaming messages – it's that impressive!

Try it now for free at the link below, no account needed. 👇

[Luiza Jarovsky, PhD @LuizaJarovsky]

🚨 AI Policy Alert: The German Federal Office for Information Security publishes the report "Generative AI Models - Opportunities and Risks for Industry and Authorities." Quotes & comments:

"LLMs are trained based on huge text corpora. The origin of these texts and their quality are generally not fully verified due to the large amount of data. Therefore, personal or copyrighted data, as well as texts with questionable, false, or discriminatory content (e.g., disinformation, propaganda, or hate messages), may be included in the training set. When generating outputs, these contents may appear in these outputs either verbatim or slightly altered (Weidinger, et al., 2022). Imbalances in the training data can also lead to biases in the model" (page 9)

-

"If individual data points are disproportionately present in the training data, there is a risk that the model cannot adequately learn the desired data distribution and, depending on the extent, tends to produce repetitive, one-sided, or incoherent outputs (known as model collapse). It is expected that this problem will increasingly occur in the future, as LLM-generated data becomes more available on the internet and is used to train new LLMs (Shumailov, et al., 2023). This could lead to self-reinforcing effects, which is particularly critical in cases where texts with abuse potential have been generated, or when a bias in text data becomes entrenched. This happens, for example, as more and more relevant texts are produced and used again for training new models, which in turn generate a multitude of texts (Bender, et al., 2021)." (page 10)

-

"The high linguistic quality of the model outputs, combined with user-friendly access via APIs and the enormous flexibility of responses from currently popular LLMs, makes it easier for criminals to misuse the models for a targeted generation of misinformation (De Angelis, et al., 2023), propaganda texts, hate messages, product reviews, or posts for social media."

➡️ According to the report, special attention should be given to the following aspects:
➵ Raising awareness of users;
➵ Testing;
➵ Handling sensitive data;
➵ Establishing transparency;
➵ Auditing of inputs and outputs;
➵ Paying attention to (indirect) prompt injections;
➵ Selection and management of training data;
➵ Developing practical expertise.

➡️ Of the dozens of AI reports published lately, this one is especially detailed regarding AI-related risk and potential countermeasures.

➡️The document is a must-read for people developing AI or working on AI policymaking and regulation, especially pages 8-28.

➡️ Link to the @BSI_Bund report below.

➡️ For more information on AI policy and regulation, subscribe to my weekly newsletter (link in bio).

[Qredence.ai @QredenceAI]

One of the reason of why some of our solution will be open-source , only after it's actually developped. Open-source license will be defined without any futur change for it.
Long project vision since start. Put Ethics as a core value for real, without needing to scream it.

[Zachary BENSALEM @ZacharyByDesign]

Yep
an example of it.
Liked the use of the term of "rug pulled" from @t3dotgg
https://t.co/nDwWblRfek

[Zachary BENSALEM @ZacharyByDesign]

Oh and also another thing

[Zachary BENSALEM @ZacharyByDesign]

More infos later this week, but basicly, the way you will be able to interact with the provided model will slighly change
Spoiler : won't be just some context window conversation
Spoiler 2 : will be incremently added
Spoiler 3 : your data will never be used to train any internal LLM

[Qredence.ai @QredenceAI]

👇🚀

[Zachary BENSALEM @ZacharyByDesign]

Changelog for @QredenceUI , updates, current features and upcoming ones .
👇
#1 GPT-4-Turbo is free to use for the whole weekend on https://t.co/nmytzYhIcM !

[Zachary BENSALEM @ZacharyByDesign]

Yes i recommend ! Especially for implementing a basic context window into a chat !
However I don't see streaming except maybe on https://t.co/nmytzYhane 👀🥰

[Zachary BENSALEM @ZacharyByDesign]

We added Claude 3 Sonet and Opus on @QredenceAI btw ! Sonet is free to use for now ! Built using @FlowiseAI

[Zachary BENSALEM @ZacharyByDesign]

📡 gpt-4-0125-preview is no longer free, Basic subscription are at 5$/month to get access, also to the upcoming models like Mistral-Large, and few surprises.
We do not use your data to train upcoming custom models
@QredenceUI will still provide free access to gpt-3.5-turbo and few others, all you need is create an account.

Built natively in @weweb_io , get streaming text generated and the always up-to-date LLMs.
More details on the unified experience cross model very soon too.

#buildinpublic

[Microsoft @Microsoft]

How can you implement a winning AI strategy across your organization?

Our new research and five-element framework highlight a few best practices to consider: https://t.co/NppU2DHrWt

[Qredence.ai @QredenceAI]

Currently adding GPT-4-Turbo to subscriber
then Mistral-Large !

An account will be needed for the free models in order to avoid abuse !

Starting at 5$/month to access premium ones in a unified interface !
https://t.co/qVPuuXGJEJ

[Zachary BENSALEM @ZacharyByDesign]

Sneak peak of what to expect in the coming days.
Streaming chat native in @weweb_io
Adding GPT-4-Turbo

...
And @MistralAI Large !
More infos tomorow

#buildinpublic

[Qredence.ai @QredenceAI]

A template ? With streaming ? like on a vercel template but easily customizable through @weweb_io to you without writing one command on your terminal ? 👀
A discount code for all the Qredence subscribers ? 👀
See you all on monday, it's still the beginning 🚀

[Zachary BENSALEM @ZacharyByDesign]

Back on track very soon ! Will make a stealth initial launch of the new interfaces, and on monday videos and few things, the UI library is planned to be send around tuesday, i will do it by batch (in order to not send each time a new request to @weweb_io !

[Qredence.ai @QredenceAI]

👇

[Qredence - AgenticFleet @AgenticFleet]

The new landing page is up !
https://t.co/HGCRJ9OOM0

Few things to adjust but focus on the webapp now !