boognish_bear said:

---

🚨BREAKING: Microsoft Research + Salesforce just dropped a paper that should scare every AI builder.

They tested 15 top LLMs GPT-4.1, Gemini 2.5 Pro, Claude 3.7 Sonnet, o3, DeepSeek R1, Llama 4 across 200,000+ simulated conversations.

Single-turn prompt: 90% performance.… pic.twitter.com/lhbQHA2OPb

— Hasan Toor (@hasantoxr) February 18, 2026

---

So don't question the AI, or you will greatly increase the possibility of a bad result. Nice!

whiterock said:

---

heard a stat yesterday. Takes 1.4Gig of power to run 1Gig of data.

the human brain can run on a ham sandwich & a 10oz bag of Lays potato chips. It's a stronger server, but tends to get distracted by repetitive number crunching.

---

So, sandwich munching and chips crunching takes less power and is more enjoyable than using a 'puter to crunch numbers.

boognish_bear said:

---

What kind of democracy is this https://t.co/ZhHiLarVBD

— Ryan Grim (@ryangrim) February 18, 2026

---

If it was an oil company Super PAC-ing anti environmental legislation no one would bat an eye.

canoso said:

---

A very wise friend says we human beings create idols to do our bidding and inevitably end up doing theirs.

---

Not to be mean, but neither Siri nor Alexa really count as friends.

Oldbear83 said:

---

canoso said:

---

A very wise friend says we human beings create idols to do our bidding and inevitably end up doing theirs.

---

Not to be mean, but neither Siri nor Alexa really count as friends.

---

Nice... You are on a roll this week. TRT shots kicking in?

FLBear5630 said:

---

Oldbear83 said:

---

canoso said:

---

A very wise friend says we human beings create idols to do our bidding and inevitably end up doing theirs.

---

Not to be mean, but neither Siri nor Alexa really count as friends.

---

Nice... You are on a roll this week. TRT shots kicking in?

---

I don't have that issue, FLBear, but thanks I guess.

Oldbear83 said:

---

FLBear5630 said:

---

Oldbear83 said:

---

canoso said:

---

A very wise friend says we human beings create idols to do our bidding and inevitably end up doing theirs.

---

Not to be mean, but neither Siri nor Alexa really count as friends.

---

Nice... You are on a roll this week. TRT shots kicking in?

---

I don't have that issue, FLBear, but thanks I guess.

---

No problem, you do what you can.

I don't think most people grasp how dangerous this idiot is, or the power he yields. It's idiots in charge like this that got us in the mess we're in.

— Preceptitude (@preceptitude) February 21, 2026

Oldbear83 said:

---

canoso said:

---

A very wise friend says we human beings create idols to do our bidding and inevitably end up doing theirs.

---

Not to be mean, but neither Siri nor Alexa really count as friends.

---

Hilarious. I never speak to either one of "them." As in "ever." Even funnier, you yourself have experienced, as have we all, exactly what my friend so wisely says.

canoso said:

---

Oldbear83 said:

---

canoso said:

---

A very wise friend says we human beings create idols to do our bidding and inevitably end up doing theirs.

---

Not to be mean, but neither Siri nor Alexa really count as friends.

---

Hilarious. I never speak to either one of "them." As in "ever." Even funnier, you yourself have experienced, as have we all, exactly what my friend so wisely says.

---

I know, I was making a joke. An observation on how some people may like a machine better than real people.

Interesting to see FLBear's response.

hodedofome said:

---

My thought tonight is that nobody will go into finance because AI is going to do it. Since I am in finance, and AI is not going to take it over as quickly as everyone thinks, I'll be the last one standing who can do it.

And I will charge a lot of money to do it.

---

Funny you mention that. I am an old credit & collections manager (sometimes Credit on the front end, sometimes chasing payments on the back end.

I have seen a growing number of companies use AI as an excuse to cut headcount, on the assumption that the automated process could do the same work as paid employees, maybe better.

Sometimes the AI companies are promising that (never in print, always verbal, of course), and sometimes the employer jumps to that conclusion.

Not to burst your AI bubble which is making everything from electricity to consumer electronics more expensive for all of us, but here's Elon Musk predicting fully self-driving Teslas by next year for ten years straight: https://t.co/emCLtcspV3 pic.twitter.com/nn96cXj4dI

— Daractenus (@Daractenus) February 21, 2026

boognish_bear said:

---

Not to burst your AI bubble which is making everything from electricity to consumer electronics more expensive for all of us, but here's Elon Musk predicting fully self-driving Teslas by next year for ten years straight: https://t.co/emCLtcspV3 pic.twitter.com/nn96cXj4dI

— Daractenus (@Daractenus) February 21, 2026

---

Can they make left turns yet?

When are they gonna come up with Artificial Wisdom? Wait..........they already have.

boognish_bear said:

---

Not to burst your AI bubble which is making everything from electricity to consumer electronics more expensive for all of us, but here's Elon Musk predicting fully self-driving Teslas by next year for ten years straight: https://t.co/emCLtcspV3 pic.twitter.com/nn96cXj4dI

— Daractenus (@Daractenus) February 21, 2026

---

mind-numbing neo-Luddite thinking....

when demand outstrips supply (thereby driving up price), the proper response is to ramp up supply (to lower prices).

Finally, someone talking sense. AI is made of algorithms, which are structured sets of instructions. This makes it extremely effective at recognizing patterns and producing outputs based on statistical learning. But it is tiresome to watch people insist that it is somehow more… https://t.co/p9LY3HuHZG

— Hans Mahncke (@HansMahncke) February 21, 2026

boognish_bear said:

---

Finally, someone talking sense. AI is made of algorithms, which are structured sets of instructions. This makes it extremely effective at recognizing patterns and producing outputs based on statistical learning. But it is tiresome to watch people insist that it is somehow more… https://t.co/p9LY3HuHZG

— Hans Mahncke (@HansMahncke) February 21, 2026

---

Yes!!!

Guess we will wait and see if this was a knee jerk

It’s official:

IBM stock, $IBM, just posted its worst day since October 2000 after Anthropic announced that Claude can streamline COBOL code.

Today is the day AI became dystopian for millions of people. pic.twitter.com/OHaVWbZrdT

— The Kobeissi Letter (@KobeissiLetter) February 23, 2026

Anthropic's CEO Explains His Refusal to Back Down to the Pentagon.

Amodei explained his deep concerns over "autonomous drone swarms" and mass surveillance.

He pointed out a crucial reality: our military's constitutional protections rely entirely on human soldiers having the… pic.twitter.com/IsmbKecwsB

— Wes Roth (@WesRoth) February 24, 2026

boognish_bear said:

---

Finally, someone talking sense. AI is made of algorithms, which are structured sets of instructions. This makes it extremely effective at recognizing patterns and producing outputs based on statistical learning. But it is tiresome to watch people insist that it is somehow more… https://t.co/p9LY3HuHZG

— Hans Mahncke (@HansMahncke) February 21, 2026

---

I agree with his assessment of most Ai today. Where he misses is on AGI, which in no way can mimic the human adaptive logic function, but does write its own algorithms, and with competing agentic logic loops does create its own form of independent thinking. Due to computing limitations, it works well in silos. But we're still quite far from the conceptual potential some have represented.

People are mostly hung up on LLMs like ChatGPT, Gemini or Claude, which were never designed to solve broad based complicated problems. Only to better organize and frame known data.

boognish_bear said:

---

Anthropic's CEO Explains His Refusal to Back Down to the Pentagon.

Amodei explained his deep concerns over "autonomous drone swarms" and mass surveillance.

He pointed out a crucial reality: our military's constitutional protections rely entirely on human soldiers having the… pic.twitter.com/IsmbKecwsB

— Wes Roth (@WesRoth) February 24, 2026

---

I understand the principle, but you'd be hard pressed to be useful to a defense department without "kill" options in your outputs.

ATL Bear said:

---

boognish_bear said:

---

Finally, someone talking sense. AI is made of algorithms, which are structured sets of instructions. This makes it extremely effective at recognizing patterns and producing outputs based on statistical learning. But it is tiresome to watch people insist that it is somehow more… https://t.co/p9LY3HuHZG

— Hans Mahncke (@HansMahncke) February 21, 2026

---

I agree with his assessment of most Ai today. Where he misses is on AGI, which in no way can mimic the human adaptive logic function, but does write its own algorithms, and with competing agentic logic loops does create its own form of independent thinking. Due to computing limitations, it works well in silos. But we're still quite far from the conceptual potential some have represented.

People are mostly hung up on LLMs like ChatGPT, Gemini or Claude, which were never designed to solve broad based complicated problems. Only to better organize and frame known data.

---

But people are using it for analysis. If I hear one more time Co-pilot can do that cheaper, no it cannot. It can sort or reorder existing data but it can not do analysis or create. These programs are going to be used for more than designed, outcomes are dicy.

Think about this for a second...

A random guy accidentally hacked into 7,000 robot vacuums. Live camera and audio feeds inside people's homes.

Now think about every other device in your house.

Your TV has a microphone. Your phone has two cameras that are always on. Your smart… https://t.co/x1pqLhfkph

— Bark (@barkmeta) February 24, 2026

This Anthropic and Pentagon standoff is not getting enough attention. Anthropic is prepared to walk away from a $200M contract because they’re concerned about how the Pentagon wants to use their tech for autonomous weapons and mass surveillance. We are at a dangerous intersection

— Karly Kingsley (@karlykingsley) February 24, 2026

AI weapons that can't disobey illegal orders is the scariest sentence I've read this year

— Clément (@en_Clement) February 24, 2026

boognish_bear said:

---

This Anthropic and Pentagon standoff is not getting enough attention. Anthropic is prepared to walk away from a $200M contract because they’re concerned about how the Pentagon wants to use their tech for autonomous weapons and mass surveillance. We are at a dangerous intersection

— Karly Kingsley (@karlykingsley) February 24, 2026

---

Yup. There are a whole lot of AI issues that no one wants to discuss. CEOs are racing to implement, higher bonuses which is what they live for, without any regard for impacts.

FLBear5630 said:

---

boognish_bear said:

---

This Anthropic and Pentagon standoff is not getting enough attention. Anthropic is prepared to walk away from a $200M contract because they’re concerned about how the Pentagon wants to use their tech for autonomous weapons and mass surveillance. We are at a dangerous intersection

— Karly Kingsley (@karlykingsley) February 24, 2026

---

Yup. There are a whole lot of AI issues that no one wants to discuss. CEOs are racing to implement, higher bonuses which is what they live for, without any regard for impacts.

---

If the reporting is accurate credit to this CEO for at least tapping the brakes and not jumping in with the US defense since he has some concerns.

Stark County, Ohio

Yesterday, Canton City Council announced an increase in utility bills for all customers starting next month, citing 'necessary infrastructure upgrades and increased operational costs.'

They failed to mention the 5,600-square-foot, 1.5 MW redundant data… pic.twitter.com/VLLTy3br0p

— The Constitutionalist 🇺🇸 (@WeWillBeFree24) February 24, 2026

🚨Harrodsburg, Kentucky residents rally to stop a proposed AI data center they say could drain up to 110 million gallons of water a year.

Farmers warn it would squeeze wells and production, leaving locals to pay the price for someone else’s servers.pic.twitter.com/p8s8F3vdMg

— Derrick Evans (@DerrickEvans4WV) February 24, 2026

BREAKING: Amazon, Google, Meta, Microsoft, xAI, Oracle and OpenAI are set to sign an agreement with the Trump Administration to generate their own electricity supply for AI data centers, per Fox News.

This includes:

1. Meeting with President Trump in March to sign the pledge…

— The Kobeissi Letter (@KobeissiLetter) February 25, 2026

Scoop: Pentagon takes first step toward blacklisting Anthropic https://t.co/WSRdM0cTbr

— Axios (@axios) February 25, 2026