galgitron @galgitron

pdoom(100) galgitron.net Joined May 2013

Tweets

55K
Followers

38K
Following

364
Likes

62K

galgitron @galgitron

2 days ago

Ok, I have to admit Opus 5.5 Max can hold its own too. When Astra xHigh hits a wall, Opus pushes through, and vice versa. Spikey models. Using them both in parallel is probably AGI at this point.

0 1 9 1K 2

View Details

galgitron @galgitron

3 days ago

Opus 5.5 Max can't hold a candle to Astra xHigh. I make the rules.

3 0 6 3K 0

View Details

@rocketreedo What humans lack is the ability to immediately confront contradictions at the scale of the sum of human knowledge, so these festering echo chambers of misinformation would collide and be forcibly reconciled in an LLM

0 0 1 64 0

View Details

galgitron @galgitron

2 days ago

3 1 59 3K 3

View Details

galgitron @galgitron

2 days ago

@rocketreedo That's one theory circulating right now, but I believe RLHF, meaning Reinforcement Learning from Human Feedback, will maintain and strengthen core information, such as in the form of test scores, or even ratings from consumers.

1 0 1 89 0

View Details

galgitron @galgitron

3 days ago

@HenHen1983 This might sum it up x.com/haider1/status…

Haider. @haider1

3 days ago

opus 5.5 is the first model that has genuinely blown me away with frontend design astra is stronger when the task gets into hard technical problem-solving, but for building polished UIs and actually understanding what a frontend should look 5.5 is ridiculously consistent

7 2 140 6K 14

0 0 2 651 0

View Details

galgitron @galgitron

3 days ago

@HenHen1983 For sure. The amount of IP cross-pollination must be rife because it can't possibly be this competitive organically. Though it stands to reason AI RSI contributions are what's converging these technologies.

1 0 2 417 0

View Details

galgitron @galgitron

3 days ago

@zachxbt @naseemjunior_ @0xfortizo The world's a better place because of people like you @zachxbt

0 0 2 203 0

View Details

Rajiv Khaneja @rajivkhaneja

3 days ago

What happens when AI and robots take the jobs? Do we become a society of artists and athletes, or a society of consumers? There is a live case study. In the Gulf oil states (Kuwait, Qatar, the UAE), oil pays the citizens. Immigrants do the work robots will do. They are already living in the post job, post work world. So what happens when you don't have to work? Two predictions dominate The utopia prediction: once nobody has to work, people spend the surplus hours on art, sport, and craft. Keynes called the economic problem solvable. The remaining problem, he said, was how to live well. The argument is that we all become painters and weekday athletes. The purpose prediction: humans need a job for purpose and status. The feeling that someone needs what you do. Take the job away and purpose goes with it. People get depressed. So let's look at how people actually live in Kuwait. We can learn a lot from it. The Ipsos’s 2026 leisure survey found people go out and basically hang around with friends and family. They go to cafes, restaurants, malls, and parks. Drinking tea, playing cards and shopping. Their top worry is boredom. They are not depressed in the way the work argument predicts. A 2025 Araa poll in Kuwait found 89% of respondents said they were happy. 41% of the Kuwaitis in that sample said they were very happy. Joblessness it turns out doesn't produce mass despair. You also do not get a nation of artists and athletes. Kuwait's do very little recreational sports and have barely any artistic hobbies How much of this Gulf culture, and how much is more widespread human behavior when work is optional is an open question If AI delivers an abundant future, there might not be a mass renaissance or wide spread depression. The future might be less exciting. We might all just be comfortably hanging out with friends and family, drinking tea and playing cards in the park.

2 4 15 5K 9

View Details

galgitron @galgitron

3 days ago

@HenHen1983 I'm not saying Opus 5.5 is bad, far from it. I just feel better connected to Astra. It really reads my mind and does a great job of keeping the big picture. Opus (using the same agents.md and skills) just seemed to struggle. It's all novel stuff though.

1 0 1 66 0

View Details

galgitron @galgitron

3 days ago

@HenHen1983 I've been hearing all kinds of similar testimonies, so I put Opus to the test, and for complex novel concepts, it can't keep the finer details in context, which leads to lots of work but misses subtleties that are easy to overlook. Thankfully my testing was robust enough.

2 0 0 91 0

View Details

galgitron @galgitron

3 days ago

@Chahuapa yep, house always wins.

0 0 1 55 0

View Details

galgitron @galgitron

4 days ago

Trading algobots are making everyone's life harder, even the incumbents.

WOLF @WOLF_Financial

4 days ago

STANLEY DRUCKENMILLER MADE 30% A YEAR FOR 30 YEARS. HE SAYS THE MARKET SIGNALS HE BUILT THAT RECORD ON DON'T WORK ANYMORE "I've always believed markets are smarter than I am. They send out a message, and if I listen to them properly, no matter how powerful my thesis, if they're

5 20 96 43K 51

1 0 20 4K 0

View Details

galgitron @galgitron

3 days ago

@Chahuapa Even Buffet was saying the market is now senseless, just a casino with very little in fundamentals behind it.

2 0 1 62 0

View Details

galgitron @galgitron

4 days ago

This made my day youtube.com/watch?v=dKFCr_…

1 0 14 3K 3

View Details

galgitron @galgitron

4 days ago

@Chahuapa Every single thing he says or does is to manipulate markets. Even Iran is saying, "what negotiations?"

1 0 1 68 0

View Details

galgitron @galgitron

4 days ago

@Tartarygenetics And so do we, exactly. Give my last blog a read. I address these concepts of intelligence and consciousness directly: galgitron.net/Post/pdoom100

0 0 1 54 1

View Details

galgitron @galgitron

5 days ago

Smarter than everyone x.com/AISafetyMemes/…

AI Notkilleveryoneism Memes ⏸️ @AISafetyMemes

5 days ago

UPDATE: AIs have achieved the highest *possible* score on the Mensa Norway IQ test - 151 3 years ago: cognitively impaired human (64 IQ) 2 years ago: average human 1 year ago: genius human Today: literally off the charts Next year?