A.I. is Progressing Faster Than You Think!

[sdv]

Posted by ThePythonicCow (here)
While listening to Nate Jones analyze the upcoming approx trillion dollar IPO's (initial public offerings) for two of the largest A.I. companies, Anthropic and OpenAI in his latest video at 'OpenAI Just Filed For Its IPO. The Real Story Isn't The Trillion Dollars, it occurred to me that we're seeing something like what my grandfather saw, when electricity went from the laboratory, to a deeply integrated world-wide infrastructure, a century ago.

Then it was power going "BIG". Now it's intelligence going "BIG".

The electric power grid did not replace all power. The intelligence grid will not replace all intelligence. Both were, or will be, expanded and refined in many ways, opening new opportunities for (back then) industrially produced power or (coming to us now) industrially produced intelligence, including down to small devices like say a solar powered "smart" watch ... where the industrial scale is on the back end, making millions of the same small device.

As I posted to Nate's above video:

So Anthropic and OpenAI become like the Westinghouse and General Electric of electrical equipment, a century ago. Destined to become huge while the underlying technology was advancing quickly, but then having to compete with many other smaller, sometimes much smaller, companies specializing in particular refinements of the technology and use cases.

In short: Machine power (aka electricity) == Machine intelligence (aka A.I.).

If neither psychopaths with nukes nor a galactic plasma current wave take down our civilization, one might be able to predict the shape of A.I. development over coming decades by comparing it with the shape of electricity development, a century ago.

That is an interesting way of looking at the development of AI that I had not thought of before.

I go way back to the printing machine for how new technology changed the world. Not only was knowledge (and entertainment) made available to all, but if you could cobble together the funds, you could print your own pamphlet and then distribute it. Along with enlightenment, disinformation spread like wildfire, but what an opportunity for advancing civilization!

[bojancan]

We know AI is a tool... but, how can you be pro..or anti tool.... at the end, the tool is either useful or it's not.... as I understand... but at the end... at the end.. not to useful to anyone... my 2 cents...
Interesting talk......

June, 15
CRASH IMMINENT: Ed Zitron Says AI Valuations Are Complete FRAUDS

Source: https://youtube.com/watch?v=jDGPnaX-86Y

[/YOUTUBE]

[ThePythonicCow]

Posted by bojancan (here)
We know AI is a tool... but at the end... not to useful to anyone... my 2 cents...

Skilsaws might be dangerous tools, that have removed a few limbs by accident, and that one would not let a child play with.

But many a carpenter would not want to forego his Skilsaw. I find AI to be quite useful.

[lake]

I consider the current AI's the same way that I used to consider the internet 25-30 years ago .... it could be a great boon for this world or it could be the major point of "despair"!

The internet has not become the "great boon" which I had hoped for although it still has a chance to reverse what the greater percent of it has become and hopefully that will happen? <<< I live in my own little realm of positivity

I do use AI and even pay for AI (I pay because the free versions are nothing compared to the paid ones!) and the currently available ones are quite something! And they are even getting better and better very quickly.

I am self taught at writing code and never had a "job" in code development (multi trade carpenter) but I have made many websites, forums and currently look after 2 of my own servers and 1 for a company (only because the bloke who owns the firm is a mate of mine and I like him as a being) but the way that AI creates code is amazing! And the speed which it does it quite shocking !!!!

I wrote some code back around 2008 which was for a firm to track 1000's of individual parts through multiple processes and the code was image based.
Since then I have always thought that a "game" could be made from that code but I never did anything about the consideration.

A couple of weeks ago Grok Build was released and I have been using that!
I gave it the old code, told it what I wanted and after some chat about different concepts which I wanted to achieve .... I have the game I always thought could be made from the old code

It hasn't taken weeks of work, just some of my free time (still work full time) and in truth, I have played the game for longer than it took Grok Build to make it!

It actually made an AI for the game (for people to play against) with 3 levels of difficultly and after refining that game AI quite a bit .... it plays a very strong game!

I asked Grok Build to create a file describing how the AI works and (in part ) this is what it made:

## Technical Explanation

The AI is a **hybrid negamax alpha-beta searcher** with an explicit two-action look-ahead for the first move of each turn.

### Core Architecture

- **Negamax + alpha-beta pruning** (`search` function):
A standard game-tree search. It explores possible sequences of actions, alternating between itself and the opponent. Alpha-beta pruning cuts off branches that cannot possibly affect the final choice, dramatically reducing the work.

- **Depths** (in actions, not turns):
- Easy: ~3
- Medium: ~4
- Hard: ~6

Because each "turn" gives two actions, depth 6 on Hard means it can look roughly three full turns ahead (its action, yours, its, yours, its, yours) in the main search.

- **Snapshot / Restore** (`snapshotState` / `restoreState`):
Before trying any hypothetical move it saves the entire relevant game state (all piece positions and charges, both barriers, breakthrough counts, whose turn it is, actions left). After exploring that line it restores the state exactly. This lets it reuse the real movement code without corrupting the visible game.

- **Search guards** (`aiSearching` flag):
While the AI is thinking, it temporarily replaces `log`, `render`, `playSound`, `checkWin`, `endGame`, etc. with no-ops. The actual state mutations still happen (so the rules are followed exactly), but the player sees and hears nothing until the AI has chosen its real moves.

From what I can gather, this game AI is much better that near all other game playing AI's??
Of course, I have gone through the AI code and it would have taken me years to make (IF I had ever thought of the way to do it but as said, I am not a coder!) so I rather impressed with Grok Build!

Is it good or bad that AI can write code like this and so very quickly through many files, changing each file as the new version requires from the prompt I give?

I think that it comes back to "Is the internet a force for good or bad?"

And that narrows down to "Who controls it and are they a good or bad person?"

But right now .... I have a game that I couldn't have made (or it would have took me so very long to do and I haven't that time to give it) which I have held in my mind for near 20 years .... So right now, for me, selfishly, .... I will say yes (but I retain the right to change my mind and consider it a bad thing at the drop of my hat!).

[ThePythonicCow]

Posted by lake (here)
Of course, I have gone through the AI code and it would have taken me years to make (IF I had ever thought of the way to do it but as said, I am not a coder!) so I rather impressed with Grok Build!

As a professional (now retired) programmer for almost a half century, AI is deeply trained on and now thinks like I do. In those areas where I am current and active, I take the lead. But in the 999 out of 1000 areas that I might want to venture into, AI knows that area fluently and guides and teaches me as fast as I can keep up, or would gladly just do the work for me, if I wanted. In my prime, I could stay actively skilled in a handful of areas at once, whereas AI can do that in thousands of areas at once, and at 10 or 100 times the speed.

(The above applies in particular to Fable 5, which we coders got a brief opportunity to use last week. Opus 4.8 is close, and other more general purpose chatbots have their uses, but not usually as high end coders.)

[Johnnycomelately]

Posted by ThePythonicCow (here)

Posted by lake (here)
Of course, I have gone through the AI code and it would have taken me years to make (IF I had ever thought of the way to do it but as said, I am not a coder!) so I rather impressed with Grok Build!

As a professional (now retired) programmer for almost a half century, AI is deeply trained on and now thinks like I do. In those areas where I am current and active, I take the lead. But in the 999 out of 1000 areas that I might want to venture into, AI knows that area fluently and guides and teaches me as fast as I can keep up, or would gladly just do the work for me, if I wanted. In my prime, I could stay actively skilled in a handful of areas at once, whereas AI can do that in thousands of areas at once, and at 10 or 100 times the speed.

(The above applies in particular to Fable 5, which we coders got a brief opportunity to use last week. Opus 4.8 is close, and other more general purpose chatbots have their uses, but not usually as high end coders.)

Replying to both lake, and Pythonic Paul.

Lake, your graphic is hard for me to interpret as a playboard for chess, but I want to believe that your AI helper is a mini-me of Deep Blue.

Paul, there is thunder sounding into my open window right now, so I am reminded of the fragility of electronic works. Anyways, how do you feel about the advantage, say the leverage, that coding gives you over the rest of us normie plebes? Please express your response as if you are a twelve year old dynastic prince in 3 or 4K BCE Egypt.


Now about me, haha.

Some months ago, maybe last calendar year I said I would announce it if I ever pulled da trigger on using AI. Well, that time has come.

A month-plus ago, I got sold a change in phone service. Been a hassle, lessons learned. My new provider (Bell, away from Telus) offered 12 months of free Perplexity Pro, apparently a CDN$30/mo thing. I tested a couple simple tasks, but then my deeper quiry return disappointed me. Wanted to see how my classmates in uni turned out, with their accomplishments and present states.

If anyone knows how to use that AI (Perplexity Pro) effectively, I’m all ears. I do get that quality of prompts/questions is crucial for working these bots.

[ThePythonicCow]

Posted by Johnnycomelately (here)
Anyways, how do you feel about the advantage, say the leverage, that coding gives you over the rest of us normie plebes? Please express your response as if you are a twelve year old dynastic prince in 3 or 4K BCE Egypt.

Perhaps its the same leverage my car's mechanic has over me, who can fix my car (er, eh, chariot or whatever they had in ancient Egypt) when I can't readily do so. Of more importance to me, I got paid some decent money for many years because companies needed good coders, so I had leverage to get more lucrative jobs than many workers.

Posted by Johnnycomelately (here)
12 months of free Perplexity Pro, apparently a CDN$30/mo thing. I tested a couple simple tasks, but then my deeper quiry return disappointed me.

Posted by Johnnycomelately (here)
Wanted to see how my classmates in uni turned out, with their accomplishments and present states.

I find Perplexity.ai quite useful, and it's one of my favorite AI tools ... for particular uses.

The strength of Perplexity.ai is that it is constantly searching high activity websites and keeping up to date. I observe this both in the answers I get from Perplexity, and in the Avalon server logs, where I can see who or what is scanning the public side of our forum.

For a recent example of how I used Perplexity, I was having trouble sorting through the various Amazon pages for a particular kind (specific connectors and cable lengths and such) of ear buds, that I use to listen to audio on my computer. The mostly Chinese vendors of such mass produced items do not accurately describe and delineate all the 100's of minor variations of ear buds and their cabling, and Amazon's sales and search pages don't help either. After an hour or two of not being able to be sure what I was looking at met my requirements or not (frustratingly confusing and contradictory details on each Amazon page). I finally asked Perplexity. It figured out exactly what I required, and what I preferred, and guided me to exactly the right product ... which I am now happily using.

In summary, I use Perplexity for more complex queries that it can answer from the well trodden websites that I know or suspect it tracks daily, or for the latest news it would have seen in those searches. Google has gone from being my primary search engine to being less frequently used mostly for cases where a few key words or names will bring up information that has been stable for a while, like using an encyclopedia, dictionary, or atlas.

I doubt Perplexity tracks millions of ordinary individuals across their various moves, social media posts, academic studies and employment, so I wouldn't expect Perplexity to fill me in on what happened to my school classmates, except in the case one of them turned out to be famous, or infamous.

As when dealing with shop keepers, coworkers, doctors, lawyers, mechanics, plumbers, ... one interacts with some potentially useful ones a bit to figure out what they're good at (or not so good at), and how efficient they are of your time, money, and patience.

[lake]

Posted by ThePythonicCow (here)
As a professional (now retired) programmer for almost a half century, AI is deeply trained on and now thinks like I do. In those areas where I am current and active, I take the lead. But in the 999 out of 1000 areas that I might want to venture into, AI knows that area fluently and guides and teaches me as fast as I can keep up, or would gladly just do the work for me, if I wanted. In my prime, I could stay actively skilled in a handful of areas at once, whereas AI can do that in thousands of areas at once, and at 10 or 100 times the speed.

(The above applies in particular to Fable 5, which we coders got a brief opportunity to use last week. Opus 4.8 is close, and other more general purpose chatbots have their uses, but not usually as high end coders.)

I can say that it (the advances in AI writing code) has taken a lot of the "fun" and "learning" out of the hobby I had!

As an example .... The game was designed (by me) with 7 "flows" each of which has 6 positions within the flow.
Once that was all working and tested, I considered that 7 positions per flow may be a "tougher" game?

So I copied the 6 "position" directory into a new directory, gave that file path to Grok Build and told it that I wanted to see how a 7 position game would play?

It re-wrote all the files so that it would now play as 7 positions per flow (including changing the game AI) and did it in under 5 minutes!

I have tested and played the 7 position per flow game multiple times and there is not an error anywhere which I can find!

There is nothing for me to "fix".

BTW .... the 6 position is a fast, tactical game which averages about 5-8 minutes to play .... the 7 position takes about 12 -20 minutes to play and the game AI is much stronger in the 7 position game!

[ThePythonicCow]

Posted by lake (here)
I can say that it (the advances in AI writing code) has taken a lot of the "fun" and "learning" out of the hobby I had!

...

There is nothing for me to "fix".

Maybe it is time to take on more challenging problems.

[Johnnycomelately]

Posted by ThePythonicCow (here)

Posted by lake (here)
I can say that it (the advances in AI writing code) has taken a lot of the "fun" and "learning" out of the hobby I had!

...

There is nothing for me to "fix".

Maybe it is time to take on more challenging problems.

Solid advice, for other life quandaries as well. On the other hand, mom told me that pyramids aren’t built in a day.

[Richter]

CANCELLING AI: A TRANSCRIPT
June 12, 2026
Joseph P. Farrell

As you've guessed, I've been on an artificial intelligence focus in this week's blogs, with Monday's blog about the "gnostic technique of contrived orthodox contexts" being used by an artificial intelligence to make fake diagnoses based on completely falsified "data" injected into the data stream, and with Wednesday's blog about Anthropic's co-CEO Olah speaking to Vatican officials, presumably to get the Vatican's imprimatur on Anthropic's AI in a rather more sophisticated version of the same Gnostic tactic of contriving an orthodox context for something.

But in today's blog, some artificial intelligence executives are playing the role of Nebuchadnezzar/Nabuchodonossor and are seeing the hard mathematical handwriting on the wall. This story is courtesy of L.G.L.R, and before we get to it, let me clarify why today's blog is about a video: today's blog is less about the video than about its transcript, which L.G.L.R. graciously also included in the email alerting me of the story, and included in such a way that he had clearly not simply copied and pasted the transcript, but had edited it back into a readable and acceptable form. So not only a thank you for sharing the story, but a double thank you for including a nicely edited transcript. It is the transcript to which we shall refer, and not the video. In any case, the video and more importantly, the transcript) decimate the transhumanist agendas surrounding artificial intelligence that were so confidently advanced a few years ago.

Source: https://youtube.com/watch?v=NBtUgWn-nHs

CONTINUE: https://gizadeathstar.com/2026/06/ca...-a-transcript/

[ThePythonicCow]

Posted by Richter (here)
CANCELLING AI: A TRANSCRIPT
June 12, 2026
Joseph P. Farrell

As you've guessed, I've been on an artificial intelligence focus in this week's blogs ...

The Youtube video you linked is not of Joseph P. Farrell lamenting the dangers of AI, but rather has this title and description:

Why Tech CEOs Are Quietly Cancelling Their AI Plans

Tech CEOs spent the last two years promising AI would replace workers, cut costs, and transform everything. Now they're quietly cancelling data centers, rehiring humans and admitting the math doesn't work. From Microsoft pulling back on billions in infrastructure to Starbucks killing its AI inventory system after it couldn't count milk, Uber burning through a year of AI budget in four months, and one company accidentally spending $500 million on AI tools in a single month the AI hype is hitting reality. Even Sam Altman now says he was wrong about AI replacing jobs.

===

My reply to the linked "Tech CEOs ... Cancelling ... AI Plans" video:

Yes - when any major change rolls through rapidly, one can find examples of major mistakes and retakes, on both sides of the coin. Big change provides fertile ground for Big mistakes.

Some will wish they had considered AI more quickly, and some will wish they had done so less quickly.

===

My reply to Joseph P. Farrell's laments:

AI, like books, radio, television, and the Web, provide(d) opportunities to the would-be tyrants for greater surveillance and control.

They also provide(d) opportunities to would-be truth seekers for sharing insights and coordinating their efforts.

Information and intelligence tools expand information and intelligence opportunities for all.

[Squareinthecircle]

[QUOTE=ThePythonicCow;1718112]

Posted by Richter (here)
CANCELLING AI: A TRANSCRIPT
June 12, 2026
Joseph P. Farrell

As you've guessed, I've been on an artificial intelligence focus in this week's blogs ...

Why Tech CEOs Are Quietly Cancelling Their AI Plans

Tech CEOs spent the last two years promising AI would replace workers, cut costs, and transform everything. Now they're quietly cancelling data centers, rehiring humans and admitting the math doesn't work. From Microsoft pulling back on billions in infrastructure to Starbucks killing its AI inventory system after it couldn't count milk, Uber burning through a year of AI budget in four months, and one company accidentally spending $500 million on AI tools in a single month the AI hype is hitting reality. Even Sam Altman now says he was wrong about AI replacing jobs.

...

Why didn't the AIs see that coming? Did the math change? I would consider that a black mark on their recognized abilities.

IDK if this has been posted yet but thought it was worth a gander:

https://substack.com/@youcantmakethi...ction&r=2ld5cy

[ExomatrixTV]

• AI is Changing Creativity Forever (15 Disturbing Examples):

Source: https://youtube.com/watch?v=ZZst_psfeWw

• 00:00 How AI is changing creativity
• 00:44 Restoring the past with AI
  
• 02:02 Recreating voices and music
• 03:39 AI translation and virtual artists
  
• 05:32 AI avatars replacing creators
  
• 07:34 When AI content hides itself
  
• 08:39 Viral AI videos and deepfakes
  
• 10:22 Bringing creators back to life
  
• 11:16 De-aging and unfinished art
  
• 12:24 Building AI music videos
• 17:13 The future of creative work