Ezana Azene MD PhD

Vascular and interventional radiologist practicing at Gundersen Health System in La Crosse, WI #iRad

[Ezana Azene MD PhD @AzeneMD]

#IRad #InterventionalRadiology #MedicarePFS #CMS

[Ezana Azene MD PhD @AzeneMD]

CY2027 Medicare PFS final rule for interventional radiology (IR), by my quick read: the conversion factor falls 2.31% (non-APM) as the temporary 2.5% boost expires. IR work RVUs are flat, and practice expense gains offset part of it (+1% facility, +4% office). Net for hospital-based IR is roughly -1.3%. SIR has yet to release its analysis.

https://t.co/CXQ2GfX5P7

[Ezana Azene MD PhD @AzeneMD]

Ezana’s cynical HOT take 🔥🤨

The frontier labs’ efforts to “slow” AI is not only about safety.

If intelligence becomes a commodity like electricity, frontier labs risk becoming interchangeable utilities while their customers capture much of the value. Keeping their best models to themselves would help them avoid that. 

I think their goal is to create a mechanism for delaying model release without risking loss of market share to other frontier labs. Restrictions that delay public access more than internal progress would preserve an advantage for the labs (and their chosen partners).

The safety concerns are legitimate. AI has real risks. However I call bullshit on safety being the whole explanation for an “AI” slowdown.

Consider two commercial incentives:

1. Make it harder for Chinese competitors to catch up.

Anthropic has accused Chinese labs of training competing models on Claude’s outputs (distillation). Withholding newer models limits that shortcut. If the Chinese are distilling models several generations behind what the labs use internally, Chinese model development will be significantly hampered, depending on how much those labs truly rely on distillation.

2. Compete against their own customers.

Imagine buying AI for drug discovery from a lab using a model five generations ahead of yours to develop competing drugs. The lab would still need capital and industry expertise, but you would be competing against your supplier using inferior tools. I think that’s the endgame envisioned by the frontier labs.

But there’s a problem.  A lab that delays model releases will lose customers to rivals that keep releasing better models. The obvious solution is for frontier labs to coordinate delays. However, such coordination violates antitrust laws.  

Anthropic CEO @DarioAmodei has proposed a narrow antitrust waiver for safety discussions. Musk, Hassabis, and Altman agree.

I wonder why???? 🤔🤔🤔🤔

Maybe it’s because a safety exemption can easily become a route to restricting public access while the labs pursue those commercial advantages using internal models that are generations ahead of what the public has. 

Am I wrong???

#AI
#AIsafety

@elonmusk
@demishassabis
@sama

[Ezana Azene MD PhD @AzeneMD]

There are far less heavy-handed ways to address these risks than banning open-source models: KYC, open-source guardrails, engage China.

KYC: Labs can substantially limit illicit distillation with the same KYC methods used in fintech. Model-routing services that provide access to U.S. frontier-model APIs, such as OpenRouter, should meet the same standard.

Guardrails: Require U.S. companies to use frontier-level open-source models only in deployments with built-in safeguards, including on-premises deployments. Require periodic audits for sensitive sectors such as finance, defense, and critical infrastructure. @NIST

Engage China: The CCP does not want open-source models banned in the United States. Nor does it want threat actors using them to create chaos. Those interests overlap. Negotiate accordingly.

#OpenSourceAI #AIGovernance #AISafety

[Ezana Azene MD PhD @AzeneMD]

https://t.co/7fW4SJX5Dh

[Ezana Azene MD PhD @AzeneMD]

In Episode 65 Dr. Dunleavy and I chat about the challenges and importance of addressing Interventional Radiology shortages in small and rural communities, like those served by Gundersen Health System.
@emplifyghs @SIRspecialists @LaCrosseWI
https://t.co/TVF1FSOrwr

[Ezana Azene MD PhD @AzeneMD]

@Jamielkey @NicoleShanahan Very easy problem to fix. Make it easy for the H1B visa recipient to move between companies. The only reason they can be paid less or forced to work more is because they can't easily switch companies like American citizens.

[Ezana Azene MD PhD @AzeneMD]

@rockypoliticss And, many of the CEOs of the largest and most successful American companies are or have been non-white and non-European: Microsoft, Alphabet, AMD, Merck, Zoom, Adobe, NVIDIA, Broadcom, PepsiCo, Walgreens, M&T Bank, Walgreens, and many more.

[Ezana Azene MD PhD @AzeneMD]

@rockypoliticss Much of USA was built on work of "imported" people. Africans: trillions of dollars in free labor. Their descendants invented blood bank, traffic signal, gas masks, refrigerated trucks. Indian Americans: MRI coils, first email. Chinese Americans: Transistor, CD, built railroads.

[Ezana Azene MD PhD @AzeneMD]

@elonmusk @elonmusk has proposed $2T in cuts from the $6.7T federal budget. But, only $1T is non-defense discretionary spending. So, for DOGE to truly impact the ballooning deficit, it must tackle entitlement spending (unlikely?) https://t.co/QqCcNhfmpL

[Ezana Azene MD PhD @AzeneMD]

@PaulMMCooper I recently discovered your podcast. Have listed through Han Dynasty. Fantastic work! Deserving of docudrama miniseries. Plan to follow on Patreon. Thank you!

[Ezana Azene MD PhD @AzeneMD]

@DavidSacks 8 USC 1158 says “Any alien who is physically present in the United States or who arrives in the United States … may apply for asylum”. That asylum request then triggers a lot of due process, during which the immigrant disappears.

[Ezana Azene MD PhD @AzeneMD]

@DavidSacks @DavidSacks it's hell on earth. Imagine being in the position of these wretched, forcibly conscripted, young Russian and Ukrainian men... heartbreaking. Thank you for being a voice of reason. https://t.co/1iWio8Pg8p

[Ezana Azene MD PhD @AzeneMD]

@a_mneff Once you’ve diagnosed the problem, you’ll be able to address it. Don’t settle for “all medical students go through it.”

[Ezana Azene MD PhD @AzeneMD]

@a_mneff Most of the advice you’re receiving is some version of “hang in there, be kind to yourself.” I agree with this advice 110%. But, it is only a first step. You need to reflect on the underlying causes. Learning style? Undiagnosed disability or anxiety? Heart not in it?

[Ezana Azene MD PhD @AzeneMD]

To learn more about the #CTRACT study, please visit
https://t.co/vnJvvPnKHw #patientcare #bloodclots

[Ezana Azene MD PhD @AzeneMD]

You can visit https://t.co/vnJvvPnKHw to learn more. #patientcare #bloodclots #clinicalresearch

[Ezana Azene MD PhD @AzeneMD]

If you’re age 18 or older, have been diagnosed with a #bloodclot and still experience leg fatigue
or swelling, you might be a fit for the #CTRACT study: https://t.co/vnJvvPnKHw

[Ezana Azene MD PhD @AzeneMD]

Carefully evaluating patients who are treated in #clinicalresearch studies is the best way for doctors to learn what the best treatments are for specific #medical conditions. Learn more here: https://t.co/vnJvvPnKHw #patientcare