TechCollect的动态

GenAI tailored to the specific requirements of community associations is the best way to protect your data and ensure compliance.

Self-Introspection: new year's resolution With the new year approaching quickly, I already wondered what goal to set as my new year's resolution. Maybe just adding a new technical skill. But watch for yourself how AI agents in a 3D scene can help interact and guide with your goals. If you really happen to have found a consistent and achievable goal as your new year's resolution (we know they are only meant to keep us pseudo-busy) and want to make it unhappen, just consult the German Angst and the Innerer Schweinehund agents to maximize being comfortable with your discomfort... --- Technical notes: AI service is implemented using Groq cloud on a rate limited developer account. If you encounter problems or have your own API key, you can use it as well. Alternatively, you can message me to provide you with an API key for testing purposes (on a self-hosted Ollama LLM server).

Have you seen GPT-4o by OpenAI? How are you planning to integrate GPT-4o into your workflow? ?? GPT-4o is a breakthrough in AI technology with omni-modal capabilities, meaning it can process and generate responses across text, audio, and visual formats. Its advanced language understanding, multilingual support, and contextual awareness set it apart, making it a versatile tool for professionals across various industries. This matters because it allows for more natural and efficient interactions, paving the way for innovative solutions and enhanced productivity. ?? Ideal Uses for GPT-4o: ? AI-Driven Customer Service: Enhances interaction quality and efficiency. ? Content Creation: Boosts creativity with advanced language understanding. ? Multilingual Communication: Supports multiple languages for global reach. ???? I am planning to leverage it in Customer Services Use cases: Integrating GPT-4o into customer service can significantly improve the quality of interactions and operational efficiency. For example: -> Automated Responses: Quickly address frequently asked questions, freeing up human agents for more complex issues. -> Personalized Recommendations: Use AI to provide tailored suggestions based on customer data, enhancing the customer experience. -> Seamless Implementation: Leverage skills in AI and Agile methodologies to ensure smooth integration, resulting in dynamic and comprehensive customer support. Peter Gostev is hands down the go-to guy for LLM insights. I found his testing methods of LLMs are like a messy kitchen experiment - unique, chaotic, and surprisingly brilliant! #AI #GPT4o #ProfessionalGrowth #CustomerService #ContentCreation Looking forward to see your experience so far with GPT-4o?

GPT--4o vs Claude Opus vs open source Llama-3. What about open source Mistral? By the way, for tasks I design in the frame of conversational or autonomous agents implementation, namely 1. text summarization or info extraction, 2. text generation, that convert from structured-to-unstructured or from unstructured-to-structured or from unstructured-to-unstructured open source Llama-3 and Mistral are excellent !!! (where the term unstructured refers to textual data analyzed exclusively by LLMs and structured refers to traditional soft engineering information, i.e. groups of fields formatted in json or XML, processed and exchanged by traditional apps but analyzed by LLMs) #chatbot #agent #genai #llm #unstructured

The efficiency of GPT-4o is quite remarkable. I very much love the Llama 3 model. It's good, it's fast and it's open source. But honestly, for most use cases GPT-4 is much cheaper to operate.

Thousands of companies went to bed last night with a terabyte of dumb, forgotten documents. Today, they are waking up to a massive "smart" database. AI is pretty incredible! I go into more detail in the attached video, but the first project I worked on when I came to the US in 2009 was a pen with an integrated camera to record your writing when filling out a form. It might sound ridiculous today, but this was the state of the art 15 years ago. Things have changed. Box, the cloud storage platform used by millions of people, just released its AI API in beta. They are sponsoring this post. Yesterday, most of the documents people stored in Box were barely useful. But today, the story is different. Here is a link for you to see what Box released: https://lnkd.in/edV-jx5N? Some of the highlights: ? You can now chat with your documents? ? Extract data from them ? Summarize them ? Fix mistakes and change those documents ? Generate derived content Here is the best part: You don't need to upload your data anywhere or have to go here or there. Box is bringing AI where your content is. Check the link above and watch the quick video I recorded for you. Honestly, I can't keep up with how fast things are changing.

Explore CloudFix's Easy, Medium, and Advanced Strategies Throwback to a special Build On Live Episode - Cost Optimization Day! This was filmed with Darko Mesaro? and Steph Gooch, and showcased CloudFix. This was streamed 9 months ago - wow how time flies. There are now more than 60 Finder/Fixers, all of which can help you reduce your AWS bill through automation, with no impact to performance! To watch the entire episode, visit: https://lnkd.in/e3rH8MAT For a free 3 month trial of CloudFix, go visit: https://lnkd.in/gqhjdNEF #aws?#finops?#costoptimization?#awsmadeeasy?#cloudfix?#AI?#optimization

It was a privilege to be on Build On Live with Darko Mesaro? and Steph Gooch, showcasing CloudFix. It's neat to think about how many Finder/Fixers we have added since then, and we keep adding more all the time! #aws #finops #costoptimization

Are you spending hours searching ?? for a 'needle' in your company’s information 'haystack'? We have an AI-powered solution for you ??! Modern companies have a wealth of information, but much of it is buried in unstructured file formats like PDFs. That’s where Vertex AI Document Search comes to the rescue ??—a powerful Google Cloud tool that uses natural language search to quickly find answers and link you to the most relevant documents in your database. Here’s what you need to get started: ? A Google Cloud project linked to an active billing account ? Editor permissions for the project ? Enabled Builder API Check out our blog ?? for the step-by-step guide by TRKKN’s Lars Nehls and set up your own Vertex AI Document Search app in minutes, transforming unstructured documents into a goldmine of insights! ?? Read our blog: https://buff.ly/41Js55j #GoogleCloud #AI #AIsearch #TRKKN

In a new blogpost for OpenProtein.AI, Tim and I describe our protein language model, PoET, and dive into some new results. Turns out that PoET reaches perplexities on heldout UniRef50 sequences that would require 500 billion parameters with a typical masked protein language model! Meanwhile, PoET only has 200 million parameters. ?? We also updated our benchmarks with the new #ProteinGymV1 for unsupervised variant effect prediction on DMS and clinical datasets with great results. PoET is almost 50% more accurate than ESM-2 on shallow MSAs and improves in/del variant effect prediction by a wide margin. More PoET applications including using PoET-derived sequence embeddings and fine-tuning to come soon! Read the blog post here: https://lnkd.in/d4ERHepH Use PoET in the OpenProtein.AI cloud platform: https://lnkd.in/dbwK2yaj PoET on github: https://lnkd.in/dxiv5KtS