essam (@itsallentropy) header
essam (@itsallentropy) profile picture

discretionary l/s generalist defy the odds

Joined July 2009
Photo by essam (@itsallentropy)Photo by essam (@itsallentropy)Photo by essam (@itsallentropy)Photo by essam (@itsallentropy)Photo by essam (@itsallentropy)Photo by essam (@itsallentropy)Photo by essam (@itsallentropy)Photo by essam (@itsallentropy)Photo by essam (@itsallentropy)Photo by essam (@itsallentropy)
9th worst week in the market ever. No crying in the casino.
19
It’s not a competition for me; DG is the most important band of the 2010s.
Get Got — Death Grips (2012)
3
112
We may be getting close to 'as good as it gets' for the dollar which means the 'debasement' trade is starting to look interesting again
1
12
4
280
24,544
Sadly I think it's safe to say that writing a 150 page essay with infographics is no longer a viable way to raise $250m with no track record
25
8
2
613
53,546
"Space exploration / Space infra" seems like an emerging narrative that will carry great beta relative to other growth narratives. Signs of high beta-shitco's you can go long are breakouts that have consolidated at a higher-lows, which describes what happened with some Space equities this year. Eyes on that factor, alongside biotech and robotics factors. Remember, you're buying positive attention, nothing more.
still think robotics and biotech narratives can catch a bid when risk returns to the market as those stories haven't been milked dry of hope yet, unlike AI stocks. Buying the dip on ai/semi's will be a death by a thousand cuts. overall indices (spy, rsp, qqq) and crypto markets have terrible price action. 30y bonds (TLT) could be a safe allocation for now. NFA.
55
Goes hard
43
667
54
27,689
693,739
As it turns out, there was no one left to buy. bid HYPE at $55
we've had trump, big Druck and Elon shill the Hyperliquid trade. I wonder where the next incremental dollar can come from? (still Long HYPE btw)
1
113
what you wanted: > curing cancer > 1 person $1b companies > productivity acceleration what you get: > all your software broken in prod > non stop cyber theft > aircraft cannot fly > "Chinese threat" Neo McCarthyism > brains too fried to even watch NFLX > surveillance state
65
319
28
3,335
173,941
As a child of the internet, it was pretty cool making internet money for the last 3 years, but I think its time to do something Real and Beautiful.
1
1
4
221
essam retweeted
> playing WoW > take a small break > check twitter > bitcoin under 69k > planes stopped working > another jensen affair > go back to playing WoW
137
749
45
13,840
253,168
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 unemploymentlabs.com/checkpo…. 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.
1,820
6,426
2,204
41,253
6,778,118
very interesting comment from an nvidia engineer i met this weekend: “CUDA isn’t the moat. Jensen is.”
35
30
15
796
95,992
On a completely empty train car back to Toronto
10
407
Sad about Jensen. Shakespearean. But as de Gaulle said, “The graveyards are filled with indispensable men.”
300
1,822
212
18,260
1,347,655
it's all really simple the next Huawei/ Deepseek collab is a cook. if it drops while ANT is trading, the stock limits down, govt can't have that. so they are going to purge everyone in favor of open source AI same reason you can't drive BYD cars. and literally same people
39
85
9
1,474
88,754
essam retweeted
New hyper niche obsession just dropped Vintage Canadian oil company annual report covers
54
331
20
5,029
151,326
still think robotics and biotech narratives can catch a bid when risk returns to the market as those stories haven't been milked dry of hope yet, unlike AI stocks. Buying the dip on ai/semi's will be a death by a thousand cuts. overall indices (spy, rsp, qqq) and crypto markets have terrible price action. 30y bonds (TLT) could be a safe allocation for now. NFA.
1
1
1
208
How it feels to be 29
189
8,223
384
125,470
1,579,876