ML Brilliance
@mlbrilliance.com
🤖 ML/AI guy turning complex algorithms into simple solutions | RPA wizard | Automation advocate | Sharing the magic of machine learning one post at a time 🧠✨ | #AIforAll
Reposted by ML Brilliance
So there's ONE actually good and useful case, which is Retrieval Augmented Generation, because it's *actually* better and faster than human brains at parsing and making connections in huge corpora. But that's almost never what GenAI true believers are talking about when they say this stuff.
There are at least four extremely good arguments against generative "AI" - any one of which would be disqualifying on its own - and I have literally never seen a single counterargument that isn't just "it's cool" or "you can't stop it."
December 27, 2024 at 1:45 AM
So there's ONE actually good and useful case, which is Retrieval Augmented Generation, because it's *actually* better and faster than human brains at parsing and making connections in huge corpora. But that's almost never what GenAI true believers are talking about when they say this stuff.
Reposted by ML Brilliance
That's a lot of parameters for a large language model.
The human brain has ~100 trillion. About 60X GPT-4.
The human brain has ~100 trillion. About 60X GPT-4.
December 22, 2024 at 4:10 PM
That's a lot of parameters for a large language model.
The human brain has ~100 trillion. About 60X GPT-4.
The human brain has ~100 trillion. About 60X GPT-4.
Reposted by ML Brilliance
Added detailed blog post on this topic to explain what each does goreplay.org/blog/load-te...
Load Testing: Core Concepts, Types, Best Practices and Tools
A comprehensive guide to understanding and implementing effective load testing, including detailed coverage of testing types, metrics, tools, and real-world examples.
goreplay.org
December 23, 2024 at 1:43 PM
Added detailed blog post on this topic to explain what each does goreplay.org/blog/load-te...
Reposted by ML Brilliance
BREAKING NEWS: Minutes ago, on the 12th day of "shipmas", @OpenAI announces their new o3 models.
o3 is much, much more than a "fancy auto-complete". :) It can reason through complicated problems. Is it Artificial General Intelligence? I don't know. But it's certainly Awesomely Great Intelligence.
o3 is much, much more than a "fancy auto-complete". :) It can reason through complicated problems. Is it Artificial General Intelligence? I don't know. But it's certainly Awesomely Great Intelligence.
agent.ai | The Professional Network for AI Agents
Get things done with the help of AI agents
Agent.ai
December 20, 2024 at 7:42 PM
BREAKING NEWS: Minutes ago, on the 12th day of "shipmas", @OpenAI announces their new o3 models.
o3 is much, much more than a "fancy auto-complete". :) It can reason through complicated problems. Is it Artificial General Intelligence? I don't know. But it's certainly Awesomely Great Intelligence.
o3 is much, much more than a "fancy auto-complete". :) It can reason through complicated problems. Is it Artificial General Intelligence? I don't know. But it's certainly Awesomely Great Intelligence.
Reposted by ML Brilliance
BERT is so back 🔥
Answer AI and Lighton released ModernBERT: lightning-fast state-of-the-art BERT model with Apache 2.0 license 🥹
2x fast as debertav3 and 3x faster than nomic 💨
all models are here hf.co/collections/answerdotai/modernbert-67627ad707a4acbf33c41deb
read more hf.co/blog/modernbert 📖
Answer AI and Lighton released ModernBERT: lightning-fast state-of-the-art BERT model with Apache 2.0 license 🥹
2x fast as debertav3 and 3x faster than nomic 💨
all models are here hf.co/collections/answerdotai/modernbert-67627ad707a4acbf33c41deb
read more hf.co/blog/modernbert 📖
December 19, 2024 at 4:55 PM
BERT is so back 🔥
Answer AI and Lighton released ModernBERT: lightning-fast state-of-the-art BERT model with Apache 2.0 license 🥹
2x fast as debertav3 and 3x faster than nomic 💨
all models are here hf.co/collections/answerdotai/modernbert-67627ad707a4acbf33c41deb
read more hf.co/blog/modernbert 📖
Answer AI and Lighton released ModernBERT: lightning-fast state-of-the-art BERT model with Apache 2.0 license 🥹
2x fast as debertav3 and 3x faster than nomic 💨
all models are here hf.co/collections/answerdotai/modernbert-67627ad707a4acbf33c41deb
read more hf.co/blog/modernbert 📖
Reposted by ML Brilliance
#econjm #econjmp
Dear Econ JM candidates
1. If applying to applied / metrics job, have ur code on github.
2. If you have #datascience / #ML skills, show it. We might want you to teach a ds for econ class.
3 Latex is not a software skill we care about. It's a -ve signal in 2024.
#econsky
Dear Econ JM candidates
1. If applying to applied / metrics job, have ur code on github.
2. If you have #datascience / #ML skills, show it. We might want you to teach a ds for econ class.
3 Latex is not a software skill we care about. It's a -ve signal in 2024.
#econsky
December 4, 2024 at 4:51 PM
#econjm #econjmp
Dear Econ JM candidates
1. If applying to applied / metrics job, have ur code on github.
2. If you have #datascience / #ML skills, show it. We might want you to teach a ds for econ class.
3 Latex is not a software skill we care about. It's a -ve signal in 2024.
#econsky
Dear Econ JM candidates
1. If applying to applied / metrics job, have ur code on github.
2. If you have #datascience / #ML skills, show it. We might want you to teach a ds for econ class.
3 Latex is not a software skill we care about. It's a -ve signal in 2024.
#econsky
Reposted by ML Brilliance
Real agents are a way off. But no reason to think they're not coming.
www.infoworld.com/article/3614...
www.infoworld.com/article/3614...
Copilot Studio Agents get a major upgrade via Microsoft 365 Copilot
Microsoft uses Ignite 2024 to set out its vision for agentic AI in business.
www.infoworld.com
December 2, 2024 at 6:12 PM
Real agents are a way off. But no reason to think they're not coming.
www.infoworld.com/article/3614...
www.infoworld.com/article/3614...
Reposted by ML Brilliance
Canadian news giants are suing OpenAI for allegedly stealing their content to train ChatGPT without compensation.
Big Tech exploits creators while sidestepping the rules. If AI can't respect copyright and builds on stolen content, it's time to regulate. www.washingtonpost.com/technology/2...
Big Tech exploits creators while sidestepping the rules. If AI can't respect copyright and builds on stolen content, it's time to regulate. www.washingtonpost.com/technology/2...
Top Canadian publishers sue OpenAI, joining AI copyright fight
The lawsuit, brought by the CBC, Globe and Mail and others, shows how the battle over copyright and AI is expanding beyond the U.S.
www.washingtonpost.com
December 2, 2024 at 6:30 PM
Canadian news giants are suing OpenAI for allegedly stealing their content to train ChatGPT without compensation.
Big Tech exploits creators while sidestepping the rules. If AI can't respect copyright and builds on stolen content, it's time to regulate. www.washingtonpost.com/technology/2...
Big Tech exploits creators while sidestepping the rules. If AI can't respect copyright and builds on stolen content, it's time to regulate. www.washingtonpost.com/technology/2...
Reposted by ML Brilliance
AI-Powered Insights: Streamlining Variable Analysis with GitHub Copilot in Visual Studio
AI-Powered Insights: Streamlining Variable Analysis with GitHub Copilot in Visual Studio - Visual Studio Blog
Discover how the new Ask GitHub Copilot feature in Visual Studio 2022 can streamline your debugging process with AI-driven insights, making error analysis quicker and easier than ever before.
devblogs.microsoft.com
December 2, 2024 at 7:24 PM
AI-Powered Insights: Streamlining Variable Analysis with GitHub Copilot in Visual Studio
Science is where BlueSky is!!!
November 30, 2024 at 10:50 PM
Science is where BlueSky is!!!
ML Engineer Roadmap for me:
November 30, 2024 at 3:56 PM
ML Engineer Roadmap for me:
Reposted by ML Brilliance
This is a 1 hour general-audience introduction to Large Language Models: the core technical component behind systems like ChatGPT, Claude, and Bard. What they are, where they are headed, comparisons and analogies to present-day operating systems, and security-related challenges. #ML #AI #LLM
[1hr Talk] Intro to Large Language Models
YouTube video by Andrej Karpathy
www.youtube.com
November 28, 2024 at 3:11 AM
Reposted by ML Brilliance
Andrew Ng released "aisuite", so we added it to observes. Start observing your AI models but then lightweight.
`pip install observers[aisuite] # or observers[litellm]`
Release:
github.com/cfahlgren1/o...
`pip install observers[aisuite] # or observers[litellm]`
Release:
github.com/cfahlgren1/o...
Release 0.1.3 - Support for `aisuite` and `litellm` · cfahlgren1/observers
What's Changed
feat: initial packaged version by @davidberenstein1957 in #2
feat: argilla support by @davidberenstein1957 in #3
add datasets example by @cfahlgren1 in #4
Improve quickstart example...
github.com
November 27, 2024 at 11:19 AM
Andrew Ng released "aisuite", so we added it to observes. Start observing your AI models but then lightweight.
`pip install observers[aisuite] # or observers[litellm]`
Release:
github.com/cfahlgren1/o...
`pip install observers[aisuite] # or observers[litellm]`
Release:
github.com/cfahlgren1/o...
I know there is some disdain about AI here but I come in peace to learn and share.
November 28, 2024 at 2:55 PM
I know there is some disdain about AI here but I come in peace to learn and share.