Hershey Unlocks Deeper Insights with SAP Business AI https://youtu.be/MKo5biKm_wk?si=d88Um_rrUa9lNYC1 via @YouTube #SAP #BusinessAI #genAI #CustomerStory #SAPChampions @sap

Galicia Transforms HR with SAP Business AI https://youtu.be/7hlXbjX1a2s?si=A1qnpJVBek8XBtTq via @YouTube #SAP #BusinessAI #CustomerStory #genAI #SAPChampions @sap

#AI #GenAI #edtech #academia

I like the linked summary of current major issues. Mike often favors a particular tech/company in his columns, but getting past that, he has stated the problems well.

The patchwork solutions are still not good enough for real buy-in by academia. (I mean, the cited 88% is maybe a B+ if you want to put it in those terms.)

The techbros need a new paradigm to connect with educators. It's not 1995 or 2005 or even 2015.

https://www.computerworld.com/article/3843220/finally-some-truth-serum-for-lying-genai-chatbots.html

Je repense à ce toot. En vrai les LLM sont un projet anti-science. Ils vont contre l'idée qu'expliquer comment on arrive à un résultat est important.

Ils habituent à tout accepter par argument d'autorité, venant d'un outil qui n'est jamais capable de retracer son raisonnement.

Les LLM promeuvent l'idée qu'il est normal d'affirmer des choses sans pouvoir les prouver, et sans aucune transparence.

https://mamot.fr/@burgervege/114138389357668883

Ah yes, allowing models to train on my content to have the ability to generate fucking stupid boring videos.

No thanks. I spare the models from the stupidity and boredom that I generate.

I get to make videos and watch them. Anything's a bonus, but they're not for training.

Grap/Graph/Graphics.. GPU

04:38am on a Friday means it's time to upgrade my local rack's LLMs

Gemma2 -> Gemma3 (1,4,12,27b)
Llama3.2 -> Llama3.3 (70b)
Phi3 -> Phi4 (14b)
DeepSeek-R1 (1.5, 7, 8, 14b, 32b, 70b)

Also some SysD/journald puke to clean up via new automation calls (have they no regard for garbage collection? no they do not).

Don't you really think its time we slowed down and put some thought into the #words we are using.

Perhaps we need to value words more, appreciate their nuances and imprecision, understand that they can can different meanings depending on how they are combined?

I'm so tired of trying to fathom rushed emails and post replies

I doubt that #GenAI use is helping

Please, can we just slow down and take some care with our use of words?

"#OpenAI wants the US government to ensure it has access to any data it wants to train #GenAI models, and to stop foreign countries from trying to enforce #copyright rules against it and other American AI firms."

https://www.theregister.com/2025/03/13/openai_data_copyright/

Beyond the Hype: Is architecture for AI even necessary?
https://blog.scottlogic.com/2025/03/13/beyond-the-hype-is-architecture-for-ai-even-necessary.html

Here’s how I use LLMs to help me write code https://simonwillison.net/2025/Mar/11/using-llms-for-code/
#llm #developers #genai #coding

Welp, #appleintelligence in the new #iPhone #iOS update has completely bollixed the dictation and #Siri on the phone. The #AI is completely mistaking commands for queries, even offering #ChatGPT to figure out what I am saying. #Apple needs to realize that if non-technical people get a bad impression of their system, they will turn it off and never turned it on again. Why I keep on turning it on, I don't know. In any case, I am turning it off, again.

#BoostingIsSharing

Technology secretary Peter Kyle asks ChatGPT for science and media advice | Peter Kyle | The Guardian
https://www.theguardian.com/politics/2025/mar/13/technlogy-secretary-peter-kyle-asks-chatgpt-for-science-and-media-advice

Incredible - and sad

Coming next week: Season 3 of the #ITOps Query #podcast, focused on #Observability in the age of #AI. #generativeAI #GenAI #o11y https://www.podbean.com/ew/pb-bjueg-183de88

Even AI tells you that you will get more stupid if you let it do everything.
#genai #ai

https://forum.cursor.com/t/cursor-told-me-i-should-learn-coding-instead-of-asking-it-to-generate-it-limit-of-800-locs/61132

How Generative A.I. Complements the #MAGA Style

Online #Trump supporters have embraced a unique form of #irony that is hard to parse — and easy to deploy with new technologies.
#ai #genai #generativeai #politics

https://www.nytimes.com/2025/03/13/magazine/generative-ai-maga-style.html

Revolutionizing #KnowledgeDiscovery with #GenAI to Transform #DocumentManagement | by Roddam Harshavardhan Reddy | #Intuit Engineering | Mar, 2025 | Medium
https://medium.com/intuit-engineering/revolutionizing-knowledge-discovery-with-genai-to-transform-document-management-0cdf4385c11c

Reading about Copilot and how to put guardrails on it. There are anecdotal reports that one of the most common questions to Copilot in organisations is, “What’s the salary of the CEO?”