53 yo cripple out here trying to use AI to make the world a better place, have discussions about new and upcoming tech, meet new folk, and make people smile
I cannot wait to be able to do more than 15 second clips and do a movie!! This is amazing so very cool. Now if only imagine had a better voice module to get noises and different voices instead of using the same generic voice
Did you know that the Terafab isn’t just rising on empty Texas dirt—it’s being built on the bones of a dead coal plant, and SpaceX already owns the river pump that will keep its massive cooling reservoir alive?
The site sits at the former Gibbons Creek Steam Electric Station (a shuttered coal-fired plant) in Grimes County near the reservoir that once cooled it. Instead of sucking local groundwater dry like a normal mega-fab, Terafab will use that reservoir for industrial water, treat wastewater on-site, and—this is the detail almost no one is posting about—SpaceX already owns the Navasota River pumping station so it can divert stormwater into the reservoir and keep the whole system topped up. They’re also building their own ultra-clean natural-gas power plants on-site so the thing never leans on the ERCOT grid.
Elon is literally turning a dirty-energy corpse into the largest, most valuable building on Earth (100 million sq ft, vertically integrated logic + memory + packaging for Optimus, Cybercabs, and orbital AI data centers). Pure industrial alchemy.
Super-fanboy bait: Is this the ultimate flex of “make the future from the ruins of the past,” or is it just Elon quietly securing a private water-and-power fortress while the rest of the semiconductor industry begs for grid connections and municipal water? Drop your take.
Most people treating their X analytics like a scoreboard are playing the wrong game.
Impressions, engagement rate, “top posts” — these numbers feel important because they’re precise. They’re also mostly noise.
A post can rack up 100k impressions and move zero needle.
A quiet post with 800 views can land the right 12 people and change your trajectory.
The algorithm rewards what keeps people scrolling, not what actually matters.
And half the “growth” tactics people obsess over are just optimizing for metrics that don’t convert into anything real.
The dirty truth: X analytics are great at telling you what the platform likes. They’re terrible at telling you what you should care about.
Stop letting the dashboard set your strategy. Use it as a flashlight, not a compass.
How many of you have watched a “low-performing” post quietly do more for your goals than the ones that lit up the analytics? Or are you still chasing the green numbers? #algorithm
@mayemusk I got an award when I was a kid .it was from president Regan for reading the most books in my area in a summer. I love books! I have no idea where that award is. Every time I asked my grandma where it was she would always tell me she didn't know. Lol
@elonmusk Man. I've been playing around with it for a few days and I've made some pretty cool looking stuff. But in order to understand the command structure in order to get that type of detail is crazy...where can I learn how to do this?
Fun facts from computer history...
In 2011, two Amazon sellers’ pricing algorithms got into a feedback loop and priced a biology textbook about fruit flies at nearly $24 million.7ebb95
WIRED
The book was The Making of a Fly: The Genetics of Animal Design by Peter Lawrence (an out-of-print developmental biology reference). Used copies were going for around $35. Two third-party sellers, “profnath” and “bordeebook,” each had a new copy listed and were using automated pricing tools that adjusted based on each other.
One algorithm was set to undercut the other slightly (roughly 0.9983 times the competitor’s price). The other was set to price itself higher (about 1.270589 times the lower price), possibly because it didn’t actually have the book in stock and wanted a margin if someone ordered it (so it could buy from the cheaper seller and still profit).
Every day the prices ratcheted upward in a perfect loop. They climbed from around a million dollars into the multimillions. On April 18, 2011, the peak hit $23,698,655.93 (plus $3.99 shipping). A UC Berkeley biologist, Michael Eisen, noticed it while trying to buy a copy for his lab, tracked the daily jumps, and reverse-engineered the multipliers.407047
WIRED
Eventually a human at one of the sellers spotted the absurdity and manually reset the price to something sane (around $100). The other algorithm promptly followed. No one bought the $23-million version, of course.
It’s a classic example of what happens when simple rules (undercut / mark up relative to the other guy) run without any sanity check or upper bound. The algorithms did exactly what they were told—perfectly, relentlessly, and hilariously. #algorithm
@toosoft221 A week ago this was the way. Today if you don't want to sit down and think about a way to create valid content that people will want to react with there is no reason to do this type of posting
@uday_devops Yes. So far. Granted most of the heavy lifting was already done by gpt. But getting simulated testing done in order to have a product ready for a human to do physical testing was a nightmare. Grok did it in 6-8 hours
As soon as the rule change happened they all flew the cookies nest. Most of the people that came in here for money never had a single genuine thought process in their brain ever. It's easy. Make original content that a high end niche of verified content creators will engage in, or simply stop waisting your and everyone else's time lol. The algorithm is king now. Bow down to it or suffer the consequences