Opus is cool but it costs 3x on premium tokens in copilot so I stick with Sonnet as much as I can haha
I have a $200 a month budget at my company, and just today I burned through $101 using Opus 4.8, lol. It was a time sensitive task I had to get done, so I put it on Max effort, haha.
"‘More harmful than helpful’: young people sour on AI" https://archive.is/newest/https://www.ft.com/content/73fc962e-ce68-4521-9c5d-841a666eed10
AI is killing the summer internship. The entry-level pipeline that built careers is breaking. Remote work -- not AI -- has sidelined recent college graduates, research finds
Yeah I’m on Ed’s newsletter list and have read a bunch of his (super detailed) posts, which are all pretty scathing of AI as a whole. His views appear to make sense, but if what he says is true, then you think the companies wouldn’t keep making the decisions they are. I’m not quite as doomer as he is, as I think Anthropic and Google, and to a lesser extent OpenAI, are going to keep growing their revenues and be able to pay whatever expenses come their way. But we’ll see.
Accenture shares fall to lowest since 2017 as AI threat mounts https://archive.is/20260618152229/h...eb-92ab-ee0f91385c62#selection-1477.0-1477.62 Its only going to get worse for all consulting firms. Soon you'll be able to get a connector/MCP for every data provider and pipe that into your AI tool, and have that do the analysis for you that you need, versus pay a consulting firm thousands/millions.
The amount of high end talent Anthropic continues to poach from OpenAI and Google is crazy Nobel laureate John Jumper is leaving DeepMind for rival Anthropic | TechCrunch
I like having Gemini built into my phone but I'd be lying if I said AI in general doesn't scare the shit out of me sub
The speed its advancing is pretty crazy; they keep learning on top of what was learned before, so each model seems like a huge step up, and these advances are happening at a faster and faster rate. The top labs have core incentives to get into our working and private lives as much as possible, to make their services as sticky as possible, so they'll continue to know more and more about us, knowing everything about the companies and people that use their services. Real, hard data. We are at a point though that the US government is now pausing model releases (first Fable, and now ChatGPT's 5.6), so we might actually see a slowdown in AI in society, at least for the time being.