Author Topic: New AI Models - Gemini 3.0, Claud 4.5 Opus, ChatGPT 5.2, Deepseek V3.2, Grok 4.2  (Read 668 times)

Palumboism

  • Getbig IV
  • ****
  • Posts: 3833
The last three weeks have been a fire storm of new AI models.  All six of the top models came out with impressive updates in that time.

Gemini 3.0

Claud Opus 4.5

ChatGPT 5.2

Deepseek V3.2

Grok 4.2

Mistral 3.0

Expect this cadence of model releases to continue going forward, which means AI will improve even more rapidly than people expected.

When Google came out with Gemini 3.0, Sam Altman, CEO of open AI, came out with a code red announcement to all employees.  That's how dramatic an improvement it was.


Zillotch

  • Getbig V
  • *****
  • Posts: 5709
  • the smoke of their torment ascendeth up for ever
The last three weeks have been a fire storm of new AI models.  All six of the top models came out with impressive updates in that time.

you are an imbecile.

Palumboism

  • Getbig IV
  • ****
  • Posts: 3833
you are an imbecile.

I won't say what I want to say.

obsidian

  • Getbig V
  • *****
  • Posts: 8263
you are an imbecile.
If insults built intelligence, you’d finally have some.

Kwon

  • Getbig V
  • *****
  • Posts: 54114
  • PRONOUNS: Ze/Zir
The last three weeks have been a fire storm of new AI models.  All six of the top models came out with impressive updates in that time.

Gemini 3.0

Claud Opus 4.5

ChatGPT 5.2

Deepseek V3.2

Grok 4.2

Mistral 3.0

Expect this cadence of model releases to continue going forward, which means AI will improve even more rapidly than people expected.

When Google came out with Gemini 3.0, Sam Altman, CEO of open AI, came out with a code red announcement to all employees.  That's how dramatic an improvement it was.

In two years, SKYNET 1.0!
Q

Zillotch

  • Getbig V
  • *****
  • Posts: 5709
  • the smoke of their torment ascendeth up for ever
I won't say what I want to say.

If insults built intelligence, you’d finally have some.

ai is anti life.

u two r on the soul train… destination hell.

have u dudes ever looked into the origins of computing?

Irongrip400

  • Getbig V
  • *****
  • Posts: 22831
  • Pan Germanism, Pax Britannica
ai is anti life.

u two r on the soul train… destination hell.

have u dudes ever looked into the origins of computing?


Explain brother.

Palumboism

  • Getbig IV
  • ****
  • Posts: 3833
ai is anti life.

u two r on the soul train… destination hell.

have u dudes ever looked into the origins of computing?

I'm an Encyclopedia of computer history, Claude Shannon is the origin.
 
What do you have against AI?  I seriously don't understand the hate. 

obsidian

  • Getbig V
  • *****
  • Posts: 8263
ai is anti life.

u two r on the soul train… destination hell.

have u dudes ever looked into the origins of computing?
I'd be all for merging with AI and robotics. Upload me so I can live in the cloud for billions of years with a superhuman physical body. I'd like AI to hurry up and figure out all the necessary tech to do so.

obsidian

  • Getbig V
  • *****
  • Posts: 8263
I'm an Encyclopedia of computer history, Claude Shannon is the origin.
 
What do you have against AI?  I seriously don't understand the hate.
They hate AI because it scales past them.  ;D

A lot of programmers seem to despise AI, especially when non-programmers use it effectively. I’ve written several genuinely useful scripts with AI, shared them on a forum, and instead of a simple “thanks,” the response was outrage — suddenly it was all about copyright, end-user agreements, and technicalities. What gets ignored is that it still took real work: trial and error, testing, iteration, and almost a full day to get the scripts working correctly. The hostility isn’t really about legality — it’s about feeling threatened. AI lowers the barrier, and some people don’t like the idea that skills they guarded for years are becoming more accessible.

Kwon

  • Getbig V
  • *****
  • Posts: 54114
  • PRONOUNS: Ze/Zir
Soon you can do full movies with ai.


This so far.

Q

Palumboism

  • Getbig IV
  • ****
  • Posts: 3833
They hate AI because it scales past them.  ;D

A lot of programmers seem to despise AI, especially when non-programmers use it effectively. I’ve written several genuinely useful scripts with AI, shared them on a forum, and instead of a simple “thanks,” the response was outrage — suddenly it was all about copyright, end-user agreements, and technicalities. What gets ignored is that it still took real work: trial and error, testing, iteration, and almost a full day to get the scripts working correctly. The hostility isn’t really about legality — it’s about feeling threatened. AI lowers the barrier, and some people don’t like the idea that skills they guarded for years are becoming more accessible.

I'm on my tenth AI program using six languages, ten different frameworks, and four different databases.  I can have a program up in running in less than an hour. 

I'm also using Cursor and Windsurf which are impressive in their own right.  If I want to make a change in functionality in a large code base, It tells me what lines need to be changed in what files asks if you want them made.  Changes that would have taken three hours to figure out and make take seconds.  You have to experience it to understand.  This is the exact reason the whole tech industry is being shaken to it core. 

Brenda Steunbeer

  • Getbig III
  • ***
  • Posts: 992
  • Huren Bumsen Nie Mehr Schuften!
Hundreds of billions of dollars are poured into everything AI related.
But it looks a bit similar to the first browsers many years ago: ultimately a few will dominate, and the rest is worth zero.

Which ones will be the winners and wich ones will be worth nothing?
Only Chinese and American AI seem to compete, the rest of the world is too far behind.

As far as I can tell companies like Open AI make their money by selling subscriptions and ads. Alphabet has shows that selling ads can be highly profitable. But those subscriptions: what if a chinese LLM will be free, of will cost only 2 dollars per month while ChatGPT asks 100 dollars or more per month?
People have adapted very quickly to cars, smartphones etc Made in China, and are very fond of temu and aliexpress

obsidian

  • Getbig V
  • *****
  • Posts: 8263
I'm on my tenth AI program using six languages, ten different frameworks, and four different databases.  I can have a program up in running in less than an hour. 

I'm also using Cursor and Windsurf which are impressive in their own right.  If I want to make a change in functionality in a large code base, It tells me what lines need to be changed in what files asks if you want them made.  Changes that would have taken three hours to figure out and make take seconds.  You have to experience it to understand.  This is the exact reason the whole tech industry is being shaken to it core.
Yeah, it's genuinely impressive. Claude tends to outperform ChatGPT when it comes to generating reliable, working scripts — it's consistently strong in that area. On the flip side, Claude still lags behind in image generation (at least as of my last check; it's pretty limited there).That's the nature of AI right now: it can disappoint you one moment and then blow you away the next.


Griffith

  • Getbig V
  • *****
  • Posts: 9929
  • .......
Is ChatGPT the best overall?

(If you have no use for programming code or image generation)

MuscleBuff

  • Getbig III
  • ***
  • Posts: 437
  • Getbig!
These all suck

MuscleBuff

  • Getbig III
  • ***
  • Posts: 437
  • Getbig!
.

Griffith

  • Getbig V
  • *****
  • Posts: 9929
  • .......
These all suck

'Claude Haiku' sounds like a purple-haired fat dyke.

Mayday

  • Getbig IV
  • ****
  • Posts: 3060
I'd be all for merging with AI and robotics. Upload me so I can live in the cloud for billions of years with a superhuman physical body. I'd like AI to hurry up and figure out all the necessary tech to do so.

There is the chance we Already are in there.

Brain moves at light speed so It makes way more sense to upload into a simulation and essentially live forever as a lifetime might be 1 real world second vs try and fly to another planet.

The odds are we create the matrix in here well before we truly live on another planet. We have VR, heads up displays, wearable haptics, neurallink vs Elon not able to land on the moon yet.

obsidian

  • Getbig V
  • *****
  • Posts: 8263
There is the chance we Already are in there.

Brain moves at light speed so It makes way more sense to upload into a simulation and essentially live forever as a lifetime might be 1 real world second vs try and fly to another planet.

The odds are we create the matrix in here well before we truly live on another planet. We have VR, heads up displays, wearable haptics, neurallink vs Elon not able to land on the moon yet.
Haha! It’s an interesting idea, and arguably one of the more realistic ways interstellar expansion could ever happen. Instead of moving physical bodies across vast distances, future civilizations might transmit minds as information. Consciousness could exist in a digital substrate, be sent at the speed of light, and instantiated into robotic or biological bodies prepared at the destination. Redundancy and backups would be trivial, much like cloud data today.

The major constraint, however, is infrastructure. You can’t transmit a mind into empty space — you first need receivers, energy sources, and bodies waiting on the other end. If relay stations or destination systems are spaced 100 light‑years apart, each expansion step still takes at least 100 years. Scaling that process across the roughly 100,000‑light‑year diameter of the Milky Way would take hundreds of thousands to millions of years if limited to light speed. Even with Star Trek–style warp travel, you’re still talking about expansion times measured in thousands of years.

That’s why wormholes, as depicted in Interstellar, are fundamentally different. They aren’t just faster travel; they’re shortcuts through spacetime itself. By connecting distant regions directly, they bypass both propulsion limits and light‑speed constraints entirely. If traversable wormholes were possible, they would dominate every other method of travel or communication.

Given these constraints, it’s plausible that advanced civilizations will explore simulated universes long before they meaningfully explore the physical galaxy. A simulation wouldn’t need to model reality in full microscopic detail everywhere at all times. It would only need to generate a consistent, convincing world from the observer’s point of view — rendering detail on demand, much like foveated rendering or level‑of‑detail systems already do.

I’ve worked in 3D modeling and rendering for decades, and even today we’re nowhere near reproducing the full fidelity of real life. In the real world, every grain of sand responds to gravity, organisms exist at every scale, and deeper inspection always reveals more structure. Real‑time graphics engines are still orders of magnitude away from reproducing that level of detail and physical richness universally. It’s been over 30 years since Wolfenstein 3D, and despite massive progress, we’re still far from true physical realism.

That said, full physical realism may not be necessary. A convincing subjective experience is a much lower bar than a perfect simulation of objective reality. We likely won’t see anything indistinguishable from base reality in our lifetimes — but long before that, we may see virtual worlds that are convincing enough for minds to inhabit, explore, and perhaps even prefer over the slow and hostile physical universe.

obsidian

  • Getbig V
  • *****
  • Posts: 8263
Is ChatGPT the best overall?

(If you have no use for programming code or image generation)
I use ChatGPT the most. I have multiple accounts, including one paid subscription. That said, in my opinion the free versions are nearly as capable, with the main limitation being that longer or ongoing chats get cut off once you hit the usage cap.

Claude AI, however, seems to outperform ChatGPT when it comes to script writing and programming. Many developers appear to prefer it, though I can’t say for certain that it’s the best option for those tasks overall.

Grok is extremely fast at image generation but noticeably slower when responding to general prompts. ChatGPT’s image generation is decent, but it can be slow or occasionally hang, requiring you to re-prompt to get the desired result. I’ve also encountered math errors and typos in ChatGPT’s outputs, so its responses definitely need to be double-checked.

Despite these shortcomings, ChatGPT is still a very useful and versatile tool.

I've used Deepseek a few times. It's all right I guess.

Palumboism

  • Getbig IV
  • ****
  • Posts: 3833

I've used Deepseek a few times. It's all right I guess.

The open source models from China and France are incredibly important.  I downloaded and run Deepseek R1 and will do the same for Deepseek V3.2 when it comes to Ollama.  If you look at the benchmarks, the open source models are not far behind the close source. 

When Deepseek R1 came out it was so devastating to tech that Meta fired 75 percent of their AI staff. Deepseek was that good compared to where they were.

The fact that any person or company can run these models on their own computer is huge.  I'm already planning building an AI computer.  There are multiple Youtube channels dedicated to building AI rigs for theses AI models.  It basically a gaming computer with multiple graphics cards, a massive power supply and a ton of memory.   

Silicon valley startups are focused on all the open sourced models because there is huge business opportunity there. 

Palumboism

  • Getbig IV
  • ****
  • Posts: 3833
I use ChatGPT the most. I have multiple accounts, including one paid subscription. That said, in my opinion the free versions are nearly as capable, with the main limitation being that longer or ongoing chats get cut off once you hit the usage cap.

Claude AI, however, seems to outperform ChatGPT when it comes to script writing and programming. Many developers appear to prefer it, though I can’t say for certain that it’s the best option for those tasks overall.

Grok is extremely fast at image generation but noticeably slower when responding to general prompts. ChatGPT’s image generation is decent, but it can be slow or occasionally hang, requiring you to re-prompt to get the desired result. I’ve also encountered math errors and typos in ChatGPT’s outputs, so its responses definitely need to be double-checked.

Despite these shortcomings, ChatGPT is still a very useful and versatile tool.

I don't have any paid subscriptions and primarily use Claude for programming.  When I get timed out, I'm done programming for the day.  I'm using all the models and I will occasionally get timed out on ChatGPT as well, so Gemini 3.0 coming out helps me. 

Like you said, each model is good at different things. 

When you look at the benchmarks, Gemini3.0 and chatGPT 5.2 caught up to the old version of Claude, which means they're all pretty good at programming now. 

I'm pretty good at programming, but AI allows me to 10X my output with languages and frameworks I've never used before.  It's actually a great teaching tool for programming as well.

Palumboism

  • Getbig IV
  • ****
  • Posts: 3833
That's the nature of AI right now: it can disappoint you one moment and then blow you away the next.

That's why I think it's laughable when people say AI will we the end of humanity.  I feel like I have to hold it's hand when it programs and give very simple defined tasks.  I catch it making mistakes all the time and have to correct it. 

Yes, it can blow me away, but also be comically bad at other times.  On the same day some Getbigger says that AI is going to end all life on earth, Claude can't get a button to work on a program with just one button.  I'm like, sorry, Claude can't end all life on earth today.  He's busy fixing the button on my one button app.  Maybe tomorrow.