




Hunter Durham
13.9K posts

@huntercdurham
Back from Bankruptcy | I tweet about babies and business. | Acquired 4 businesses for $3.8M | Christian Husband Father












I have zero doubt that we are seeing manufacturing starting to rip. Flatbed is naturally the first mode to signal it, as flatbeds haul heavy industrial raw materials. We talked about greenshoots a few weeks ago, and it was exciting, but still a bit iffy. I don’t recall ever seeing a national tender rejection chart surge like this. Flatbeds are on fire. This is epic.

Chicago PMI just printed 57.7 vs 52.1 expected! Was 54.0 previously. Anything above 50 signals expansion, and this is the highest reading in over 3 years. Serious acceleration in manufacturing activity. Chicago PMI often leads ISM Manufacturing PMI, which prints Monday.



Sir Isaac Newton had an estimated IQ of 190 - 200 and died a virgin. The IQ number is of course just a guess, but we know he was profoundly gifted. We also know he never reproduced. An IQ of 200 is 6.67 standard deviations above average. If accurately measured, there may be zero people alive today with an IQ of 200. It is not surprising that nature does not select for intelligence past a certain level. It becomes counterproductive to reproduction. I also think it is counterproductive to life in general. The few legitimate high-level geniuses I have known in my life did not reproduce. They did not create anything on their own. They generally were: • chronically negative • paralyzed on big decisions • lacked social skills • looked to others for direction Like everything else in life, there is a sweet spot for intelligence. Too little or too much is not good.


For $IREN I expect Amazon or Google to want the whole 1.4GW at Sweetwater 1 with option for Sweetwater 2... After hearing that cipher call I can sense the hunger for large amounts of power from Hyperscalers. 1.4GW for a single site in 2026 ready for 2027 is a large advantage At 1.4GW, it becomes highly profitable to use custom inference chips (like Google’s TPUs or Amazon’s Trainium/Inferentia), which are often 3–5x more efficient per token than general-purpose GPUs Excited to hear what $WULF has to say 👀



















We have an almost 6-year-old and an almost 2-year-old. My wife (who’s a very practical, even-keeled attorney) just looked over to me and said: “I’ve come to the conclusion that two kids are actually easier than one.”