Potentially huge for CV on edge devices - e.g., human activity classifiers running directly on cameras — without relying on cloud inference, enabling faster response times, lower latency, and improved privacy. Edge apps are going to explode over the next year.

Where are the #philosophers? Reading this it’s clear: AI will push the boundaries of human experience and impact #metaphysics (does AI undermine free will?), #epistemology (can AI really ’know’ something, can we rely on it?), #ethics (who is responsible for the actions of.

Does AI replace enterprise processes or roles? In this episode we dive into the question, give specific estimates of productivity gains, and highlight the (growing) importance of a #humanintheloop :

Up-front ideation is now next level - AI tools can produce new scenarios and prototypes 1000X faster than before, so this will quickly become enterprise best practice across the entire value-chain. But, as learned from the recent Apple whitepaper, the ability to implement these.

RT @wolfejosh: Apple just GaryMarcus'd LLM reasoning ability

AI tools are spectacularly productive, but ultimately get diminishing returns on demanding tasks (even making excuses for their diminishment). #HumansintheLoop (HITL) will become increasingly critical for sophisticated tasks. The Great Blending, AI + Human, is underway. We.

New corporate structure: a Biological workforce (humans) and a Digital workforce (AI agents); will classic “HR” primarily manage the biologic, and “IT” the digital?.

So the models have already passed average human IQ, which is not surprising. Questions: how to leverage these capabilties for positive change - do we get an explosion of productiviy and creativity? And as these models advance, what is the new "IQ test" for humans - what.

#ChatGPT’s advanced voice feature - incredibly smooth and accurate - blows away #Siri. It does’t integrate with my iPhone apps (calendar, reminders), but new models like ‘GPT-4o with scheduled tasks’ starts to address that gap. And the new 'memory feature' learns my preferences.

RT @DocumentingBTC: University of Wyoming posted a new video explaining Bitcoin with the Philosophy Professor Bradley Rettler https://t.co/….

Who says you can’t combine the study of #Philosophy with current trends in AI? I asked #ManusAI to generate a timeline of major philosophers since antiquity and an interactive web page that allows me to double-click on each philosopher and get a summary of their major.

OK, wow, this dramatically impacts the need for Sales Development Reps (SDRs), BizDev resources, basically all pre-sales/sales-qualifying roles and activity. AI is altering the workforce every day now.

Amidst all the chaos it is so refreshing to have somebody methodically and clearly illuminate a complex subject. Thanks @karpathy.

Brilliant essays from the inimitable and freakishly prescient #DavidFosterWallace. Dude is writing in the 90’s and just nailing 2025 culture - e.g., his essay about excessive TV exposure might as well be about AI and our performative social ethos: “The ability to combine [images

Michael needs it to be - probably more than any other enterprise or individual on the planet - but that doesn't mean he's wrong . .

I’ve read a lot of #HenriBergson and #Sartre in my time - 2 of my favorite #philosophers that seem most on point to our lived experience. But is was not until I studied #LLMs that I saw their close connection to modern #AI. They are all about the flow, taking life moment by.

OK #Sartre, I freely chose #Grok3 because it kicks ass. I benchmarked its summary of your key #philosophical tenants against #ChatGPT4o - it was organized better, syntactically tighter and encouraged deeper exploration. Yes, I anguished over the decision, but that is the choice I.

Agree, incredibly useful - the right amount of detail across the entire #LLM pipeline. I can't believe this is offered gratis from @karpathy . hyper-targeted knowledge diffusion, wow.

#Philosophy majors are required to study formal #logic - breaking down text into key components, often represented as symbols. #LLMs use a similar process - distilling text into tokens, usually represents as numbers. Enterprises then try to apply these models to unique.

Same. Reading “A Supposedly Fun Thing I’ll Never Do Again” and getting the same uncanny prescience. Smart dude.