I’ve pushed my cloaking research from ideal electromagnetic geometry into a discrete spherical design that suppresses scattering in simulation, even with its support rod included. At 10 GHz, the current model predicts −6.42 dB against the bare core, exceeding the −5 dB numerical target at the design frequency. That is a concrete computational result: the modeled structure changes how electromagnetic waves scatter from the object and reduces its strongest sampled signature. I’ve also independently checked the ideal spherical loss benchmarks and developed the mathematical bounds needed to test whether narrow scattering peaks escape our angular and frequency sampling. The engineering challenge is now sharply defined: make the physical resonators, dielectric structure and their interactions reproduce that response within the available error budget. The next step is a calibrated microwave experiment measuring the complete polarization response. We have successful simulated scattering suppression and a precise route to testing it in hardware, however physical validation and continuous all-angle certification remain ahead.
There’s are moments before you succeed, where you don’t know if you’re going to succeed. All the sudden it begins to work and nobody remembers being skeptical.
People need to remember, Starship is physically incapable of performing the tasks Musk has assigned to it.
Not only will it not carry 100 people at a time to the Moon nor Mars, there is no ship built out to attempt to certify for human rating. This alone is a process that takes years once there is a ship to test.
There have also been zero builds of tanker variants and depot variants. Their HLS tech map has yet to be started.
These articles launching today are not a final form to test in orbit for Artemis III.
So, if Starship CAN'T do any of the things Musk promises (has taken contract money for), and is mothballing Falcon/Dragon - what will the future of SpaceX look like?
Weather radar satellite in progress.
Will be the smallest product-ready satellite radar ever.
Step one to mastering the weather is mass producing the monitoring systems.
@sean_from_earth i won't lie to you, i think open source will be banned before too long after some major disaster. and when the day comes, you'll agree with me. i hope kimi and deepseek etc keep making models but keep them monitored on an api where they should be
I have a major result to share.
I’ve found conclusive forensic evidence that the top 4 Chinese AI labs—DeepSeek, Moonshot, MiniMax and Ant Group—have built their newest models on NVIDIA’s most advanced Blackwell chips, which are barred from export to China.
This sounds insane, but I discovered that almost every AI model carries a hidden signature of the GPU that trained it. When a model is first set up, the starting values of its weights are laid out in a pattern that depends on the exact GPU. Each combination of GPU and seed number produces a unique pattern—a sort of fingerprint of the hardware used to create the model. Traces of that pattern survive training into the published model.
Again, this evidence isn’t based on any insider information or leaking. It’s based on a forensic analysis of the model weights released by the companies themselves.
I searched more than six trillion combinations of chip layout and seed number and matched them to open-weights models. In doing so, I found conclusive evidence that six Chinese AI models—including DeepSeek-V4-Pro, Kimi K3, and MiniMax-M3—were initialized on NVIDIA Blackwell systems. Four of the models match Grace Blackwell, the GB200 and GB300 parts in NVIDIA’s flagship liquid-cooled rack-scale systems.
I then verified these results on real hardware. Running unmodified PyTorch with the seeds I’d recovered, a Blackwell B300 reproduced DeepSeek-V4-Flash’s untouched starting weights bit for bit, and a rented GB200 reproduced Kimi K3’s to a correlation of 0.999995. A Hopper H100 reproduced none of DeepSeek-V4-Flash’s, and changing the seed by one broke the match. (DeepSeek’s seed is 953049583 by the way.) The method also was verified to match 24 known open-weights models to the hardware that developers said they used.
It is exceedingly unlikely that these results are a coincidence. Across six trillion combinations, the best accidental match should reach only about 8 sigma (i.e., eight standard deviations above what random chance would produce). The real matches range from 49 to 1,385 sigma, and even the weakest single test hits 20.
These were data-center-scale training runs. With some analysis of release schedules and model details, I estimate that the largest, Kimi K3 and DeepSeek-V4-Pro, used 2,000 to 3,000 Blackwell GPUs each. Clusters of this size cost on the order of $150 to $200 million each to buy, and across these labs alone would add up to more than half a billion dollars of export-controlled GPUs, not counting experimental compute or otherwise. It’s quite likely that we see multi-billion dollar expenditure on export-controlled GPUs.
I am highly confident in this result, and it has been independently verified by senior researchers at OpenAI and Anthropic. You can read the full report at https://t.co/YOtzND0Jqk.
There is much more I’ve identified about these models through other forms of checkpoint forensics like this, and will have more to share soon.
A British P8 maritime ISR/subhunter aircraft spent about 4 hours today doing tracks above 458m deep Landsorts-abyss, just outside Stockholm.
It is a convenient spot to hide a submarine and also close to key naval installations.
To every veteran who raised their right hand, and to the families who stood behind them: thank you. Today we also remember the four-legged teammates who walked point beside them. We owe them all our best work. 🇺🇸🐾 #VeteransDay
The commitment Marines make to one another carries beyond their time in uniform. At PDW, that commitment lives on through Marines like Gage Barajas, who served as a Sergeant and small unmanned aerial systems operator and now helps support the capabilities Marines take into the field.
“The reason I wanted to join PDW was to continue to enable the brothers I had in the Marine Corps, and PDW was the obvious way to do that.”
Happy 251st birthday to the United States Marine Corps and to every Marine who carries its legacy forward. Semper Fidelis.
A team of 1,000 Fable 5.2 agents fully formalized Fermat’s Last Theorem in 30 hours, translating the proof into a language a computer can check step by step.
With the same setup, Opus 5.5 completed only 26% in 18 hours, and barely made any more progress after 30 hours.
i went to bed and asked claude to paint an autumn tree overnight.
at around 4am, it realised that it didn't have some shades -- so it stopped and titrated its own paint: diluting red, sienna and madder step by step on a second palette, then measuring each pocket from a photo.
woke up to this beautiful watercolor painting. 303 brush marks, took about 4.5 hours 🍂 (still working on the speed)
There should be a startup award for the founder who went an entire year without announcing anything.
No pivots.
No partnerships.
No groundbreaking milestones.
Just quietly sold a product and paid the employees.
Unfortunately, nobody in the industry would know who to nominate.
I fought the RL and the RL won.
I wanted a damn powerloop! It learned one powerloop, followed by three weird flip-around maneuvers. Ten days of my life that I won't be getting back.
The RL does as it pleases I guess.
Onwards.