It’s fascinating to consider how many ways Democrats are screwed from Sunday. They are looking at desolation.
* Approval poll down to 27%
* DOGE is slashing government jobs which are mostly held by Democrats.
* DOGE is slashing funding which mostly goes to Democrat-dominated organizations.
* Hollywood is falling apart due to woke films losing gobs of money, strikes, streaming, YouTube and the LA fires.
* Universities are hit by declining demographics, DOGE scrutiny, increasing skepticism of college degrees, increasing disapproval of wokeness, and AI eating into education at all levels.
* AI is going to hit white-collar jobs much harder than blue-collar.
Democrats are only beginning to realize that they are fighting for their lives and everything they know is wrong.
huxley, my daughter is working on adapting AI for use in business development at her employer. Her latest comment was that she is getting tired of AI just making stuff up. She has to call in subject matter experts to verify details. She works for a government contractor and has security clearances, so she hasn’t told me what program she’s using. Skynet, perhaps? 🙂
Every time I see one of the open thread videos, I remember to be grateful that neo is not posting her conversations with Grok.
Definitely the LLMs are going to put some people out of work, but some bloggers seemly oddly eager to put themselves out of work.
they made ai creative without being ethical, it should be the fourth law of robotics,
(how do you train a program not to lie, or exagerate, or make things up out of whole cloth,)
something many of the dystopian tales of run away computers never really considered,
maybe asimov traipsed around it in his robot detective stories,
character escapes me,
@miguel cervantes: They can’t train an LLM on “truth” because LLMs have no window into the real world, they just have text and images and people giving them thumbs up and thumbs down. They are trained to tell us either something plausible, or else today’s Preferred Narrative. Sometimes what is plausible is true, of course, and sometimes today’s Preferred Narrative is true.
Asimov’s robots had their own sense perceptions; they had to know what was real in order to obey their laws. And quite a lot of his stories involved what you might call “loopholes” in the laws of robotics. If a robot kills someone in an Asimov story, it was done in some non-obvious way for some non-obvious reason and usually at the direction of a culpable human.
His robots would lie in some cases, and in fact one of his short stories was called “Liar!” about a robot that habitually did so, seeking to obey the laws of robotics better. That robot had senses other robots didn’t, you see.
By the end of his career he had allowed his robots to discover they could use an end to justify a means, but he portrayed that as a positive development: that because they could now break a few eggs and make some tasty omelettes, robots could care for humanity much more effectively. Not all his readers agreed…
understood, even phillip K dicks stories like ‘do android sleep’ which was unfilmable until scott turned it into a futuristic noir, then, we are living in that future, the robots were programmed to be amoral, in order to do their tasks,
it’s an interesting question, which we don’t have the answer to, sadly, well much of humanity has dispensed with ethics, how can you get machines to learn,
thats a broad brush statement, but consider how two of the leading tech giants zuckerberg and gates, came to hold their position in the market
the first with the appropriation of a certain patents, the second by ‘borrowing’ code from a colleague,
I remember the golden age of blogging when all of those on the side board, were still with us, along with others like steven de beste, norm geras who was an ethical lefty, was that long ago,
they referred to this anecdote in the film about assange, I thought it couldn’t be true,
then you see how the Modern Press operates both in England and in this Country,
huxley, my daughter is working on adapting AI for use in business development at her employer. Her latest comment was that she is getting tired of AI just making stuff up. She has to call in subject matter experts to verify details.
Kate:
True. Hallucinations are a problem. Trust but verify.
Sam Altman, CEO of OpenAI, once apologized for current AIs, noting that they will be worst you will ever use.
There are now benchmark tests measuring an AI’s tendency to hallucination. For instance:
More recent AIs hallucinate less than earlier versions.
Furthermore, we are now seeing AIs like ChatGPT 4.o that include “Deep Research” options which means that the AI “thinks” for 5-30 minutes, while interrogating the web extensively for information and checking its conclusions. This reduces hallucinations as well.
“Deep Research” produces a dynamite, formatted report if that’s what you’re looking for.
…she hasn’t told me what program she’s using. Skynet, perhaps? 😉
Kate:
Don’t laugh. There is a Skynet program. It’s Chinese. It uses AI with big data analytics and facial recognition.
bibi is someone I have enormous respect for, ever since he was debating as UN Ambassador on venues like nightline, which seemed respectable, if you didn’t know any better, it turns out he was often out gunned 2-1, not only with his named rival, often hana ashwari, but apparently from peter jennings who had strong sympathies for the palestinians, and that’s being charitable and that was some 40 years ago,
then I found out about his combat experience and more recently his family’s long lineage in Zionist circle, going back too generations,
Lex Fridmen recently released a very interesting podcast with 2 extremely knowledgable AI professionals. Fair warning though, it quite long and often gets very technical. They do cover a lot of topics, but they spend some time focusing on China (specifically Deepseek and the current state of US-China relations visa vie the various trade restrictions and their ramifications). They also talk a fair amount about hardware (Nvidia) and AI ethics. Personally I found it fascinating, but I do have a bit of technical knowledge about these topics and I’m uncertain how interesting it might be to a normal person who isn’t already immersed in the stuff.
huxley, thanks for the link on hallucinating LLMs. And of course it would be the CCP which is developing a Skynet surveillance system.
don’t know that illegals pay employment and income taxes
don’t know that illegals file income tax returns and the IRS keeps track of them
think that mandatory E-Verify would somehow prevent illegals from being hired
think that the employers who hired those illegals could somehow be punished
There’d be a lot more progress if people relied less on their impression of how things should work, versus how they actually work. The laws are not what we think they are, and our misperceptions have been carefully fostered.
Pressler devoted a year of his life to turning PA red for Trump. Kudos! BTW, Pressler is a gay man who has fought against electoral fraud and sharia as well. All-around good guy.
Megyn Kelly & Co. did an episode on the recent Dem wins:
They see it as something to keep an eye on, especially since Trump won’t be on the ballot in 2026. Also, Trump isn’t addressing the economy as strongly as voters would like.
So, yes, don’t get cocky.
However, I’ve got my one eye on the longer game that the Democrats are going to lose a hell of a lot of funding and many Democrats and Deep State types are going to lose their jobs in the next two years.
There’s a huge financial crunch coming for them as their comfortable gravy trains are disrupted. Plus a lot of big Dem donors feel quite burned from 2024.
Since I see Democrats more as grifters than zealots, I suspect that they are going to be in serious disarray by the time the 2026 elections get here.
there is something to be said for the propriety and elegance of that period, were there excesses flaws no doubt,
https://twitchy.com/dougp/2025/03/28/awkward-cnbcs-joe-kernen-throws-a-wrench-in-dem-senators-selective-calls-for-accountability-n2410603
thats a more selective audience
https://x.com/megynkelly/status/1905436234865877001
It’s fascinating to consider how many ways Democrats are screwed from Sunday. They are looking at desolation.
* Approval poll down to 27%
* DOGE is slashing government jobs which are mostly held by Democrats.
* DOGE is slashing funding which mostly goes to Democrat-dominated organizations.
* Hollywood is falling apart due to woke films losing gobs of money, strikes, streaming, YouTube and the LA fires.
* Universities are hit by declining demographics, DOGE scrutiny, increasing skepticism of college degrees, increasing disapproval of wokeness, and AI eating into education at all levels.
* AI is going to hit white-collar jobs much harder than blue-collar.
Democrats are only beginning to realize that they are fighting for their lives and everything they know is wrong.
huxley, my daughter is working on adapting AI for use in business development at her employer. Her latest comment was that she is getting tired of AI just making stuff up. She has to call in subject matter experts to verify details. She works for a government contractor and has security clearances, so she hasn’t told me what program she’s using. Skynet, perhaps? 🙂
Every time I see one of the open thread videos, I remember to be grateful that neo is not posting her conversations with Grok.
Definitely the LLMs are going to put some people out of work, but some bloggers seemly oddly eager to put themselves out of work.
they made ai creative without being ethical, it should be the fourth law of robotics,
(how do you train a program not to lie, or exagerate, or make things up out of whole cloth,)
something many of the dystopian tales of run away computers never really considered,
maybe asimov traipsed around it in his robot detective stories,
character escapes me,
@miguel cervantes: They can’t train an LLM on “truth” because LLMs have no window into the real world, they just have text and images and people giving them thumbs up and thumbs down. They are trained to tell us either something plausible, or else today’s Preferred Narrative. Sometimes what is plausible is true, of course, and sometimes today’s Preferred Narrative is true.
Asimov’s robots had their own sense perceptions; they had to know what was real in order to obey their laws. And quite a lot of his stories involved what you might call “loopholes” in the laws of robotics. If a robot kills someone in an Asimov story, it was done in some non-obvious way for some non-obvious reason and usually at the direction of a culpable human.
His robots would lie in some cases, and in fact one of his short stories was called “Liar!” about a robot that habitually did so, seeking to obey the laws of robotics better. That robot had senses other robots didn’t, you see.
By the end of his career he had allowed his robots to discover they could use an end to justify a means, but he portrayed that as a positive development: that because they could now break a few eggs and make some tasty omelettes, robots could care for humanity much more effectively. Not all his readers agreed…
They are looking at desolation.
* Approval poll down to 27%
And yet, there’s this:
https://x.com/ScottPresler/status/1905355837390102721
understood, even phillip K dicks stories like ‘do android sleep’ which was unfilmable until scott turned it into a futuristic noir, then, we are living in that future, the robots were programmed to be amoral, in order to do their tasks,
it’s an interesting question, which we don’t have the answer to, sadly, well much of humanity has dispensed with ethics, how can you get machines to learn,
thats a broad brush statement, but consider how two of the leading tech giants zuckerberg and gates, came to hold their position in the market
the first with the appropriation of a certain patents, the second by ‘borrowing’ code from a colleague,
I remember the golden age of blogging when all of those on the side board, were still with us, along with others like steven de beste, norm geras who was an ethical lefty, was that long ago,
they referred to this anecdote in the film about assange, I thought it couldn’t be true,
https://archives.blog.parliament.uk/2021/03/29/john-wilkes-parliament-and-the-freedom-of-the-press-1771-2021/
then you see how the Modern Press operates both in England and in this Country,
huxley, my daughter is working on adapting AI for use in business development at her employer. Her latest comment was that she is getting tired of AI just making stuff up. She has to call in subject matter experts to verify details.
Kate:
True. Hallucinations are a problem. Trust but verify.
Sam Altman, CEO of OpenAI, once apologized for current AIs, noting that they will be worst you will ever use.
There are now benchmark tests measuring an AI’s tendency to hallucination. For instance:
–“A Ranking & Evaluation Framework For LLM Hallucinations”
https://www.galileo.ai/hallucinationindex
More recent AIs hallucinate less than earlier versions.
Furthermore, we are now seeing AIs like ChatGPT 4.o that include “Deep Research” options which means that the AI “thinks” for 5-30 minutes, while interrogating the web extensively for information and checking its conclusions. This reduces hallucinations as well.
“Deep Research” produces a dynamite, formatted report if that’s what you’re looking for.
…she hasn’t told me what program she’s using. Skynet, perhaps? 😉
Kate:
Don’t laugh. There is a Skynet program. It’s Chinese. It uses AI with big data analytics and facial recognition.
Brrr!
–“Skynet Project in China – The Rise of Ubiquitous Surveillance”
https://medium.com/@davidsehyeonbaek/skynet-project-in-china-the-rise-of-ubiquitous-surveillance-33fee53c350e#:~:text=In%20recent%20years%2C%20China%20has,time
now is it hallucinating, inadvertent, or out right deceiving, therein lies the rub,
https://x.com/Geiger_Capital/status/1905591976876990670
with this fellow I would say the latter,
Bibi (with subtitles), a first interview on the war from his perspective with Gadi Taub, 1:09:00: https://youtu.be/31bG_tEtRPE
interesting also that info would be classified,
https://www.ynetnews.com/article/s100l11rma1x
bibi is someone I have enormous respect for, ever since he was debating as UN Ambassador on venues like nightline, which seemed respectable, if you didn’t know any better, it turns out he was often out gunned 2-1, not only with his named rival, often hana ashwari, but apparently from peter jennings who had strong sympathies for the palestinians, and that’s being charitable and that was some 40 years ago,
then I found out about his combat experience and more recently his family’s long lineage in Zionist circle, going back too generations,
Lex Fridmen recently released a very interesting podcast with 2 extremely knowledgable AI professionals. Fair warning though, it quite long and often gets very technical. They do cover a lot of topics, but they spend some time focusing on China (specifically Deepseek and the current state of US-China relations visa vie the various trade restrictions and their ramifications). They also talk a fair amount about hardware (Nvidia) and AI ethics. Personally I found it fascinating, but I do have a bit of technical knowledge about these topics and I’m uncertain how interesting it might be to a normal person who isn’t already immersed in the stuff.
huxley, thanks for the link on hallucinating LLMs. And of course it would be the CCP which is developing a Skynet surveillance system.
Over at Instapundit most commenters still:
don’t know that illegals pay employment and income taxes
don’t know that illegals file income tax returns and the IRS keeps track of them
think that mandatory E-Verify would somehow prevent illegals from being hired
think that the employers who hired those illegals could somehow be punished
There’d be a lot more progress if people relied less on their impression of how things should work, versus how they actually work. The laws are not what we think they are, and our misperceptions have been carefully fostered.
To be sure, some green card and student visa holders have been getting a rapid education lately on the difference between our impression of what the laws are and what they actually are.
Re: Scott Pressler
Mike Plaiss:
Pressler devoted a year of his life to turning PA red for Trump. Kudos! BTW, Pressler is a gay man who has fought against electoral fraud and sharia as well. All-around good guy.
Megyn Kelly & Co. did an episode on the recent Dem wins:
–“Dems Winning Special Elections Now… Trouble For GOP in 2026? With Halperin, Spicer, Turrentine”
https://www.youtube.com/watch?v=5XhNrT_c8YY
They see it as something to keep an eye on, especially since Trump won’t be on the ballot in 2026. Also, Trump isn’t addressing the economy as strongly as voters would like.
So, yes, don’t get cocky.
However, I’ve got my one eye on the longer game that the Democrats are going to lose a hell of a lot of funding and many Democrats and Deep State types are going to lose their jobs in the next two years.
There’s a huge financial crunch coming for them as their comfortable gravy trains are disrupted. Plus a lot of big Dem donors feel quite burned from 2024.
Since I see Democrats more as grifters than zealots, I suspect that they are going to be in serious disarray by the time the 2026 elections get here.
meanwhile on another front
https://x.com/mypetjawa/status/1905725481661706737
https://x.com/mcgmouton57/status/1905740623916957708
and another,
the former concerns the continued detention of khalil, the latter a eo to reverse the cultural vandalism of the monuments,
Part 2: Elon Musk Fox News interview: We must make sure America will be strong for a long time
https://commoncts.blogspot.com/2025/03/part-2-elon-musk-fox-news-interview-we.html