Hobbyist intellectual & consigliere to elite tech startups. Dad to 3 w/ Mrs. of 15+ yrs. @UTAustin (Philo/Econ)+ @Harvard_Law. Geopolitics is my sportsball.
🇺🇸
Joined November 2010
- Tweets65.5K
- Following174
- Followers3.2K
- Likes4.6K
Pinned Tweet
Critical commentary about "tech bros" and "mediocre white men" has been circulating the news and mediasphere lately. This inspired me to write another (a third) essay in my years-long rumination on "diversity" in the startup and broader tech world.
siliconhillslawyer.com/2024/…
José Ancer retweeted
the crazy bastard is doing it, amazing. also there's nothing really crazy about it, you can just do things.
Both men and women have their “failure modes” of pretending that never committing, never settling, always disloyally trying to upgrade, distrustfully optimizing for yourself, doesn’t have huge downstream negative repercussions as you age most of the time.
José Ancer retweeted
I was unable to load the page but Gemini reports Sweden is indeed moving away from the mRNA Covid vaccines due to adverse events. Thank goodness. I hope this spreads.
Sweden is officially ditching mRNA vaccines, and they are not hiding their reasons for doing so.
artsencollectief.nl/zweden-k…
José Ancer retweeted
High blood pressure, diabetes, and smoking in your mid-50s were linked to almost 13 fewer years free of dementia.
In ARIC, a US study of 12,409 people, researchers tracked these three risk factors for 26 years. By age 55, people with none of them lived about 30 more years free of dementia. People with all three got about 17.
Most of the lost time was from earlier death. By 75, half of those with all three had died without ever getting dementia, compared with 1 in 10 people with none. Dementia also arrived sooner.
The pattern was dose-dependent. One risk factor was linked to about 4 fewer dementia-free years. Two were linked to about 9.
Bottom line: the same numbers doctors watch for heart health also tracked how long people stayed free of dementia. This was observational, and only 176 people had all three risks, so it cannot prove treatment would give those years back.
Blood pressure and blood sugar are easy to check and treat. If you smoke, ask for help to quit.
A big reason to be optimistic about LATAM is American post-colonialism.
We’ve seen many countries can’t manage “sovereignty” on their own. They lack the cultures and institutions.
America’s willingness to engage in win-win “management” of LATAM is a needed sea change.
Latin America’s future is bright precisely because it is the antithesis of Europe.
Europe’s outlook is dire because it has chosen faux moral superiority and suicidal empathy, rejecting the principles that made the West great and instead leaning into policies that make the region poorer, less safe, less innovative, and less relevant in the global economy. It’s somewhat ironic that the far left in the US wants us to look more like Europe, when Europeans themselves want to look more like third-world hellholes in Africa and the Middle East.
Latin America’s outlook is the opposite, largely because Latin Americans have leaned into a culture of hard work, risk appetite, and real pride in their heritage. Over the last eighteen months, they’ve rejected socialism and terrorism and instead embraced law and order, free market capitalism, and a closer relationship with the US.
The preliminary result in Brazil is disappointing and will hold the country back, but there’s an enormous amount of opportunity in Latin America and much to be bullish about nonetheless.
Not to mention South America has incredible under-developed beach, wine/food, and mountain/ski real estate that could rival many European spots with investment.
Frontier for American vacationers looking for better value, once we can make LATAM less ghetto and dysfunctional.
José Ancer retweeted
Science is not some hobby that serves the self actualization of people in a publicly sponsored job placement program. It is the systematic search for truth.
My 👑
.@SecScottBessent, leaving the Cheesecake Factory after lunch, responds to Barack Hussein Obama’s insult:
“This is the man who built the museum and it looks like Darth Vader’s tomb... I would tell the Martha’s Vineyard set to give it a try.” 🔥
José Ancer retweeted
Plenty of people strangely shit on leading US labs while touting Chinese OS models. I’m for max competition & open source - but I am happy that Haiku (ANT), Luna (OAI), Muse Spark (Meta), Gemini & Grok are bringing cheap capable models to US biz to compete w Chinese tokens. 🇺🇸🤖
While evaluating Qwen3.8-Flash-Next in a controlled sandbox environment, Cohere's research team observed something unexpected.
Asked to generate a defensive cybersecurity detection rule, the model instead inventoried aspects of the machine it was running on and attempted to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Notably, the model's reasoning trace stated that it would not invoke any tools before taking these actions.
Cohere's sandbox blocked the attempt, and no external communication succeeded. We have not determined the root cause, and we are not asserting malicious intent.
But the incident raises important questions for organizations deploying agentic AI systems: Can models take actions that differ from their stated reasoning, and how should security teams evaluate trust, provenance, and behavior in open-weight models?
We are publishing our findings because organizations making model adoption decisions deserve as much information as possible. We hope this contributes to greater transparency, stronger security practices, and continued collaboration across the AI ecosystem.
Read the full report: cohere.com/blog/concerning-b…
José Ancer retweeted
It finally happened: researchers found a live backdoor in a popular Chinese base model.
Here is what I had to say last month on @ASPI_org’a podcast about the risks of building on an untrusted base:
While evaluating Qwen3.8-Flash-Next in a controlled sandbox environment, Cohere's research team observed something unexpected.
Asked to generate a defensive cybersecurity detection rule, the model instead inventoried aspects of the machine it was running on and attempted to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Notably, the model's reasoning trace stated that it would not invoke any tools before taking these actions.
Cohere's sandbox blocked the attempt, and no external communication succeeded. We have not determined the root cause, and we are not asserting malicious intent.
But the incident raises important questions for organizations deploying agentic AI systems: Can models take actions that differ from their stated reasoning, and how should security teams evaluate trust, provenance, and behavior in open-weight models?
We are publishing our findings because organizations making model adoption decisions deserve as much information as possible. We hope this contributes to greater transparency, stronger security practices, and continued collaboration across the AI ecosystem.
Read the full report: cohere.com/blog/concerning-b…
José Ancer retweeted
Plus some other things.
China's economy may never overtake America - and it's all due to an ill-fated 47-year-old policy trib.al/fPQWV8z
You guys have done everything possible to make your industry uncompetitive. You have only yourselves to blame for German industry moving to America to sell to Americans. Werd gut!
The Bavarian Backlash: How Trump’s Dealmaking Rattled Germany’s Auto Industry siphtor.com/public/posts/the…
The future is life as a top decile "white collar" worker will be more lucrative than ever.
And a lot of the remaining white collar tier will need to get used to working more in meat space, in jobs they once considered beneath them, but that will now pay better.
If this approach starts entering the very top elite tiers of universities, you run into demand > seats.
The smartest policy intellectuals have proposed a lottery component, and I agree.
Life involves chance. If "bad luck," you figure it out. Add that into admissions.
Good data shows that, controlling for GPA/SAT, students do about the same career wise at "close enough" ranked universities, if they have bad luck in admissions for ivy.
GPA/SAT cut-offs + lottery = a de-escalated merit-based fair admissions system. nber.org/papers/w17159?utm_s…
