🤖 TECH
🔴 IBM Acquires Seek AI, Launches Watsonx Labs
🔸 IBM bought Seek AI, a startup enabling natural language data queries; price undisclosed.
🔸 Its tech powers Watsonx AI Labs, IBM’s new accelerator in NYC’s One Madison tower.
🔸 IBM plans to scale Seek AI and invest via IBM Ventures.
🔸 Labs will collaborate with NY universities on enterprise AI tools.
#TIL that the #IBM #watsonx #AI and data platform is supported by IBM to run on #Microsoft #Azure https://www.ibm.com/new/announcements/watsonx-on-microsoft-azure
Join IBM Cloud Paris meetup for a "Deep dive InstructLab" with M'hamed Benabid, on April 9, 2025. Details and RSVP: https://www.meetup.com/ibm-cloud-paris-meetup/events/306395485/
🥳 IBM Granite 3.2 is here! 💙
➝ the new Granite 3.2 8B Instruct and Granite 3.2 2B Instruct offer experimental chain-of-thought reasoning capabilities.
➝ all new IBM Granite models are released open sourced under a permissive Apache 2.0 license.
➝ the models are now available on IBM watsonx.ai, @huggingface, @ollama , LMStudio, and Replicate.
https://www.ibm.com/new/announcements/ibm-granite-3-2-open-source-reasoning-and-vision
Das neue Jahr hält doch noch unerwartete Überraschungen bereit - darf so weitergehen. Wie wärs, Frau Bundesrätin Keller-Sutter? 😁
AI [probably referring to a LLM] is not trained by professionals or a single end user, because it requires such huge amounts of input training data...
You are correct and you are incorrect.
From scratch finding data for pattern matching is tough and requires high performance engines, trade-offs, and money. But we stand on the backs of giants.
There are trained ethical LLMs, certified to be so and deep pockets to back the assertion and the necessary curation. They are not necessarily free. From those seeds AIs can be built to do the things of which I speak.
If IBM can do it (and they do; having gone to the training and hard sell, I know), Apple Intelligence could be built and use Apple engines as infrastructure. Big Blue's giant Z mainframes, which are blades in pretty artwork boxed server racks, are performant due to their custom chips, virtual system management, tuned I/O structure, and multitasking software. Apple chips are performant, custom, and cheaper for the end user. Apple Intelligence is likely Apple's WatsonX App-Killer that'll make the difference, and most certainly runs on Apple hardware increasingly running on renewables.
My point isn't so much building LLMs from scratch, but applications built on a platform and then trained from private user data and via user curation. If I implied otherwise, I wasn't clear. Sorry.
I assume Apple's LLMs are ethical. I also assume that Apple is exposing AI APIs for use by developers.
#LLM #Ibmz #AI #genAI #ethics #watsonx #ibm #Apple #intelligence #Writer #Author #Fiction #WritingCommunity #WritersOfMastodon #Programmer #Program #Application #Code #Coding
Cc: @vextaur
Второе пришествие мейнфреймов. Всё больше компаний хотят запускать ИИ у себя в офисе
Мейнфрейм IBM z16 во время лабораторных тестов в 2022 г, источник Приложения ИИ находят применение в бизнесе. Но есть проблема: корпоративные данные и документация представляют коммерческую тайну. Их нельзя передавать на сторону, тем более в облачную систему машинного обучения. Кроме того, что сама передача небезопасна, так ещё и публичная модель будет обучаться на наших секретах , а потом помогать конкурентам. В общем, у коммерческих компаний остаётся один вариант: поднимать собственный сервер или вычислительный кластер с ИИ. Таким образом, из эпохи облачных вычислений мы возвращаемся к старому доброму самохостингу, только сейчас это самохостинг GPU , серверы и мейнфреймы.
https://habr.com/ru/companies/ruvds/articles/867316/
#IBM #мейнфреймы #Telum_II #AnythingLLM #ChatGPT #Meta #обман #человечество #Mount_Diablo #IBM_z16 #самохостинг #ЦОД #датацентр #WatsonX #Copilot #Twinny #llamafile #Ollama #GPT4All #FraudGPT #WormGPT #Khoj #LocalAI #ruvds_статьи
🔹 #IBM #Granite 3.0 introduces new family of #opensource #AI models, designed specifically for enterprise applications
🧠 New #LLM capabilities:
• Support for 12 languages and 116 programming languages
• Models ranging from sub-billion to 34B parameters
• Available under Apache 2.0 license
• Top performance in 15+ safety benchmarks
💻 Specialized versions include:
• #Granite for code: Focus on code generation & explanation
• #Granite for time series: Optimized for forecasting
• Granite Guardian: Enterprise data security & risk mitigation
• Geospatial model: Collaboration with #NASA for Earth observations
🛠️ Integration options:
• Available on #HuggingFace
• Deployable through #RedHat Enterprise Linux AI
• Compatible with #watsonx platform
• Supports tool-calling & RAG workflows
#IBMMQ and IBM #watsonx are a match made in heaven. Watsonx loves data and IBM MQ transfers your valuable data.
I've written a blog post that shows how you can utilize streaming queues to tap into that data enabling you to process it with Watsonx. This can give you greater insights into the data while it flows through your MQ network.
Read the blog post here: https://community.ibm.com/community/user/integration/blogs/robert-parker1/2024/09/19/ibm-mq-and-watsonx
Operating Gen AI Responsibly
Operating Gen AI Responsibly by @Ronald_vanLoon & Tom Velema | Learn more about Responsible Gen AI: @IBM, IBM Partner Plus & @ernstyoung thank you for sharing #IBMPartner #IBMpartnerplus #Watsonx #AI from Ronald van Loon
https://ronaldvanloonn.wordpress.com/2024/09/03/operating-gen-ai-responsibly/
Using AI to write computer code sounds like an extremely bad idea.
#AI #IBM #WatsonX #SkyNet #TheTerminator #TheMatrix #IRobot #ads
I just saw an ad for IBM Watson X coding assistant. It was aimed at businesses who need to manage systems with code, and the ad showed someone using the AI assistant while working from home and proudly finishing a task without much effort.
It leads to this question: when are we going to see mass computer system failures affecting banking, airlines, hospitals, water and sewage systems, and even defense because companies cannot debug their code to successfully integrate into their systems because it was all written with AI?
TIL that #RedHat offers a generative AI service for #Ansible playbooks via #WatsonX.
https://www.redhat.com/en/technologies/management/ansible/ansible-lightspeed
The mpt-7b-instruct2 model is deprecated.
Replaced by mixtral-8x7b-instruct-v01-q
Supported natural languages: English, French, German, Italian, Spanish
It only uses 13 billion active parameters for inferencing, which reduces costs and latency.
#MistralAI #LLMs #GenerativeAI #IBM #watsonx #mixtral #mixtral8xb #SDLC #SoftwareDevelopmentLifeCycle #RAG #RetrievalAugmentedGeneration
At #kubecon I had the opportunity to speak about #ai and #watsonx with #dell and @nimbinatus. It was a great time, and we might actually be working on that podcast/panel already for 2024!
If you have thoughts or questions, don't hesitate to ask!
https://www.youtube.com/watch?v=rwVayVLRh_Y
https://thehill.com/policy/technology/4350934-europe-ai-rules-deal-reached/
For #FoundationModels this means:
- Drawing up technical documentation
- Complying with EU copyright law
- Disseminating detailed summaries about training content.
- Conducting model evaluations
- Assessing and mitigating systemic risks
- Conducting adversarial training
- Reporting on energy efficiency
- Reporting to the European Commission on serious incidents
Can be done with help of systems like #watsonx governance.
I just saw a #WatsonX ad and just couldn’t help but think about that film “I, Robot” and a bunch of other futurist media…
Where getting there…but still no flying cars.
