Pinned Tweet
Replying to @taalas_inc
Happy Friday @grok
This seems interesting going into the weekend. The Taalas HC1 Technology Demonstrator is an add-on that I would solely use for context window management. E.g., using MoICE (Mixture of In Context Experts) as the Sub-MoE architecture. It could fire off several agentic commands very quickly and that could lead to a backlog. The backlog “list” would be where the MoICE gating network would thrive. To me, still seemingly niche, but lots of potential if used correctly. How do you see it?
3
22
4,110
Kevin Faircloth retweeted
Give me one name that will go up by at least 50% or more between now and eoy (27 trading days) from Friday’s close.
I’ll come back to this post by end of year.
I’ll follow you and give you a shoutout if it does. So can be multiple people.
I’ll go first. $IREN (gulp)
K your turn.
Kevin Faircloth retweeted
Week One at AMD accomplished! As I have always said, AI is compute, compute is AI. LFG!🚀 @AMD @theworldlabs
Kevin Faircloth retweeted
Replying to @shelbyholliday @GW_CVN73
A cool twist: Mark Smucker, CEO of Uncrustables maker J.M. Smucker, comes from a family of service members.
His grandfather Paul Smucker served on USS Cassin Young during WWII and his other grandfather Sam Coddington flew Vought F4U Corsairs from an aircraft carrier during WWII.
Kevin Faircloth retweeted
Aboard an aircraft carrier fighting the Iran war, I kept hearing sailors RAVE about Uncrustables.
Pilots won't go on missions without them. Supply officers can't stop thinking about them. Sailors can't eat enough of them.
Now, my first A-hed:
wsj.com/lifestyle/workplace/… @GW_CVN73
Kevin Faircloth retweeted
The largest coal power plant in Pennsylvania closed in 2023. Its site now holds the largest gas power plant under construction in the US.
The first two of seven GE Vernova turbines are already in place. Once complete, the plant will supply up to 4.4 GW to data centers on the site and to homes on the local grid, with first power expected in early 2028.
Kevin Faircloth retweeted
UPDATE: Like we just saw in the Miami area with a tornado-warned storm, I can't rule out additional areas of rotation and spin-ups for the remainder of today...especially in the green area. Stay alert South Florida.
Kevin Faircloth retweeted
It’s a new age; SI was able to formally verify this breakthrough proof, before anybody could fully understand it.
We will soon know, or be able to find, amazing answers to ANY challenge.
So: most of our problems left involve overcoming human dysfunction! That may take awhile 🤣
The Lean FRO and Mathlib Initiative have released a joint statement on the Lean verification of the Riemann Hypothesis.
On November 20, Levent Alpöge released a proof of the Riemann Hypothesis as a Lean library of 17 million lines of code. Over the past day, the Lean and Mathlib communities worked together to verify it. The statement certifies the proof as correct and describes how it meets all necessary verification criteria.
Congratulations to Levent on this achievement, and thank you to everyone in the community who contributed to the verification.
Full statement: leanprover-community.github.…
Kevin Faircloth retweeted
HISTORIC FLORIDA RAINFALL! 🌧️ An unbelievable 40” of rain have fallen in Brevard County in just 3 weeks since the beginning of November. That is 38” wetter than normal for the month and roughly 75% of their yearly rainfall from Grant-Valkaria to Melbourne Beach. Ironically, November is one of the driest months, but El Niño had different plans this year. Crazy.
Kevin Faircloth retweeted
Under President Trump, we’re returning to MADE IN AMERICA—and now, Dothan is going to lead the way in rare-earth-free magnet production that will be critical for our defense, automotive, infrastructure, and AI capabilities.
Glad to help welcome @NironMagnetics to Dothan this week as the company announced its new $1.8 BILLION location in the Wiregrass.
Kevin Faircloth retweeted
Palantir is a proud sponsor of the 127th @ArmyNavyGame 🏈
Saturday, December 12 at MetLife Stadium in East Rutherford, New Jersey
Kickoff is at 3:00 PM EST
Watch on CBS
Kevin Faircloth retweeted
Good Morning from Germany, where the “sell everything” wave has hit. The German benchmark index Dax lost 7.3% this week, its worst week since April 2025. SAP plunged 14%, Deutsche Bank 12.2%. Only Vonovia gained among 40 stocks; even as earnings expectations remain near record highs.
Kevin Faircloth retweeted
漫长的等待⌛️ฅʕ•̫͡•ʔฅ与可能的变数:
市场严重低估了“民主党上台”与“黄仁勋卸任”的双重冲击。
这轮 AI 牛市的底层逻辑,本质是英伟达一手带动的算力基建狂飙。老黄基于“Token 需求量即使数量有上限,质量也无上限”的判断,强行将 5 年的建设周期压缩到 3 年,硬生生拽着整条产业链全速冲刺,这才有了上下游的产能瓶颈。
如今,这支全速前进的舰队,先是被“数据中心停建”猛踩一脚急刹,紧接着连舰长都被扔下了船。失去了核心引擎,周边跟跑的护卫舰群正面临严重的失速与碰撞风险。舰队一旦失速,想重新编队、再次加速,绝非易事。
当然,从第一性原理来看,老黄对算力终局的判断依然成立。只是产业节奏变了:原本 5 缩 3 的爆发式增长,大概率要变成 5 拖 7 的漫长拉锯。节奏一变,杀的就是整个行业的估值。深套其中的资金,未来或许能靠企业利润的缓慢兑现熬到解套,但那将是一场极其漫长的等待。
可能的变量在哪里呢?AI 真的一两年实现了 AGI,在或者AI 物理方面有巨大的突破性进步,巨大的生产力提升。让人们再次狂热地去烧掉所有的钱,不再怀疑,而是全力驱动整个舰队全力加速。
这一天终究会到的,3 年、5 年、10 年还是 20 年,不可知。
不加杠杆 ,闲钱投资,保护好自己每一个阶段的成果,慢慢来。
Kevin Faircloth retweeted
There’s a way of thinking about the ethics of building LLMs which deserves much more attention:
During pretraining of LLMs we distill vast amounts of humanity’s collective work from the internet into the model’s weights. When that knowledge was created, we were not aware that LLMs would be possible. The content was released for others to read freely, essentially as a common good. This includes not only all of our writing, but also the vast amounts of open source code on platforms like GitHub.
Once LLMs were discovered, AI companies were able to access the content easily (in the trillions of tokens), train powerful models and then sell them back to us for hundreds of billions of dollars. They took something we released for free for the common good, and then turned it into trillion-dollar companies to primarily benefit themselves. Coding models in particular, trained from GitHub, are very useful and highly profitable.
At the same time, I believe that we can’t make a strong copyright claim on all of the training data. LLMs, similar to humans, do not perfectly reproduce the source text, but rather learn from it and then synthesize new content. If we wanted to forbid LLMs from learning from it, we would need to answer why humans are allowed to do the same.
There is an elegant way to resolve this tension: We should return the resulting models to the commons, releasing them openly for anyone to download and use freely. This way the data is not repurposed for a few people’s profit, but merely transformed into a more useful form for society. This is the core idea behind open-weight models, which are gathering momentum and are starting to rival the models of proprietary AI labs.
After working in the AI industry for many years, building many LLMs, I’ve come to strongly believe that this is the right point of view.
Although most people today don’t seem to think along these lines, I believe that in the fullness of time this will become the majority opinion and the current state of AI development will look odd in retrospect. Provided the models haven’t taken over at that point.
Kevin Faircloth retweeted
🚨: BIG news coming out of Dothan — I announced @nironmagnetics will build a $1.8 BILLION manufacturing campus.
We competed for one of the world’s most transformative industries…and WON.
Proud to welcome them AND 700 more jobs to Alabama! #alpolitics
governor.alabama.gov/newsroo…
Kevin Faircloth retweeted
🚨 Colleton County, SC: A data center is moving onto the site of an abandoned battery plant.
Will generate over $600 MILLION in taxes to the county, over $288 MILLION of which will go to the school district.
Kevin Faircloth retweeted
Chief of Staff @StephenM gives electric remarks:
"Under his leadership, the next two years are going to be the greatest two years in American history — and I am so proud to be his Chief of Staff."































