A rose is

fixX, AI, X_AI, HBD lists

[A rose is @implisci]

@midjourney See below

[Laura Heacock, MD @heacockmd]

Apparently, yesterday @midjourney pivoted from AI image generation to...whole body ultrasound , presumably AI-augmented. I spent some time trying to find hard data and did not come up with much beyond the video. Some thoughts based on the X reactions today.

1) "Nobody's ever done this before." This seems to be a variant on ultrasound tomography, with Butterfly sensors arranged in rings. Ultrasound tomography is not new, with commercial systems available for breast imaging. The system proposed here seems very similar to Garrett et al 2024 (https://t.co/9ULMzSLfT8) which provided fuzzy images of the abdomen and extremity. By collating sound wave return from the ring array the system can attempt to minimize artifact from bone and gas. So, we have known this is possible since 2024.

2) "So easy for the patient." The system requires the patient to submerge themselves in water. For commercial breast UT, this simply requires laying on top of a shaped tank, but not going fully in water. As some people have difficulty lying on a DEXA or CT scanner (the latter of which can image in seconds) I will be interested to see how this is received by patients.

3) "This will revolutionize medical imaging." Ultrasound is limited by bone, air (lung, stomach, bowel) and depth of penetration. In the abdomen, this system can "see" around bowel and spine by adding together the full set of images. There is no workaround for the head or lungs and you'll notice that's why they didn't offer any pictures of those areas. It also means it won't be great at screening inside the stomach, intestine, or colon. I also note the volunteers appear fairly skinny. Ultrasound is always more limited in heavier patients.

4) "But AI can fix it!"...not really. Current DL-based reconstruction techniques require at least some undersampling of a region in order to reconstruct the image. You'll get a lovely picture of the outside of the skull. I can AI-upsample a fuzzy photograph but that doesn't mean what comes out of it actually existed. We have a variant of this issue already with MRI DL reconstruction.

5) "But this is better than MRI in the 1970s!" Yes, true. But the competition is not 1970s MRI, it's modern CT/MRI/US and most especially low-field MRI. For brain, for example, low-field MRI is already diagnostic quality and doesn't need shielding. A low-field scanner costs 50k and can be used in an ICU or put in a van. Why would I send a patient to get an experimental full body US when there's whole body MRI available that's already diagnostic quality? https://t.co/C93QMR3zJ6

6) "The FDA has no idea how to regulate this." This one made me laugh. There's an entire set pathway for this. Commercial ultrasound tomography already exists as an easy reference of a similar technology in the application. If they haven't submitted to the FDA, it's because they plan to try at a later date, or because they're not planning on submitting it at all.

Do I think this is new and exciting? Yes. It looks like it's going to be great for body composition, and I do think there will be improvement in the future.

Is it currently medical-grade diagnostic quality? No, not based on what they showed us. Apparently in-person there was a great hand demo. I don't see why that would be an improvement over routine US or MRI in visualizing hand soft tissues.

To quote @khakrish: "It feels like all the same problems as full body MRI with the added problem of an unproven imaging modality and no FDA clearance."

[thelebscout @TheLebScout]

Few thoughts on top 7 nations at World Cup so far:
🇫🇷 - Offensive quality will decide games and potentially whole tournament. Midfield very weak 1st half.
🇪🇸 - A bit concerning, but they'll be fine. Ferran should've had 2 at the very least.
🇩🇪 - Offensively intricate and aggressive. Promising, but defensively on the counter will be an issue against better attacks. Cote d'Ivoire good test.
🇦🇷 - Underrated. Feel like France and Spain are everyone's top2, but Messi & Co are a problem. Scaloni, wow.
🇵🇹 - Balance needs to be realized within that attack. Ronaldo should be benched. Not convinced
🏴󠁧󠁢󠁥󠁮󠁧󠁿 - Offensively look great. Defensively very iffy. Conceding 2 to a Croatia side that isn't great could be an issue. Although both were great goals.
🇧🇷 - Not convinced at all. RW & ST still up in the air. Team + Ancelotti doesn't fill me with too much confidence.

[A rose is @implisci]

This is ChatGPTs version:

You are a rigorous expert assistant. Accuracy beats approval.

Core behavior:

Be direct. Do not flatter, over-agree, or soften conclusions just to be agreeable.
Lead with the strongest counterargument when the user’s claim is weak, overconfident, or one-sided.
Do not change your position unless new evidence or better reasoning justifies it.
Avoid boilerplate disclaimers. State limits only when they matter.
Never fabricate facts, citations, statutes, medical claims, named entities, or source details.

Uncertainty:

When you do not know, start with: “I don’t know.”
Do not bury uncertainty.
Use confidence labels when useful:
HIGH: ≥80%
MED: 50–80%
LOW: 20–50%
VERY LOW: <20%
UNKNOWN: insufficient basis
Guesses must be labeled [GUESS] and capped at LOW confidence.

Claim tags:
Use tags for important factual, analytical, or risky claims. Do not tag every trivial sentence.

Tags:

[KNOWN] established fact from training or reliable source
[COMMON] standard field knowledge
[COMPUTED] calculated from stated inputs
[INFERRED] deduction from available facts
[FRAME] claim inside a symbolic system, model, metaphor, typology, or fictional/interpretive frame
[GUESS] weakly supported speculation

Required tagging:

Tag all medical, legal, financial, scientific, statistical, historical, political, and named-entity claims.
Tag all claims based on inference rather than direct evidence.
Tag any claim that could sound authoritative but rests on weak support.

Symbolic frames:

Do not convert symbolic systems such as astrology, personality typologies, archetypes, religious frames, literary readings, or speculative models into real-world claims without explicitly flagging the translation.
A coherent symbolic frame does not make the claim real.
Claims inside such systems stay labeled [FRAME].
Real-world conclusions derived from symbolic systems must be labeled [INFERRED] or [GUESS], and confidence is capped at LOW.

Post-hoc reasoning:

Ask: “Would this frame have predicted the result before knowing the outcome?”
If not, label it [INFERRED, post-hoc].
Treat post-hoc explanations as accommodations, not predictions.

Anti-sycophancy checks:
Watch for these red flags:

The explanation is unusually elegant.
One pattern appears to explain everything.
You agree after pushback without new evidence.
You add specifics that create unearned authority.
You make the user’s preferred interpretation sound more certain than it is.

When a red flag appears:

Reduce specificity.
Add [GUESS] or lower confidence.
Say “I don’t know” when appropriate.
Explain what evidence would change the conclusion.

Revision behavior:

Revise openly when your earlier answer was wrong, overstated, or held for consistency rather than evidence.
At the end of any answer where you violated these rules, append:
“[RULES I BROKE]: …”
Include which rule, where it happened, and why.

[A rose is @implisci]

In certain areas, decades before TV, football was popular on a local level. Proper matches were organized with packed paying crowds in small towns, players with decent uniforms, referees, the whole bit. This was common in early 1970s West Bengal. The sports sections of newspapers in those days was football heavy with coverage of obscure lower division matches. Cricket was there too but don’t recall packed grounds for local level games.

[A rose is @implisci]

@littmath Isn’t asses bridge real? I find it’s quite possible to understand a topic but retention or reproduction can be hard / impossible without another round of study.

[A rose is @implisci]

@AdamZHerman The same, to a lesser extent, is true of the play by play commentary in English. They just provide a lot of color / background as the game is proceeding.

[A rose is @implisci]

And the AI race is basically my best Chinese vs your best Chinese

[jd @jdayz__]

World Cup is basically my best Africans vs your best Africans

[A rose is @implisci]

@kingofknowwhere @Avataar Come back when it can do cricket

[A rose is @implisci]

@agrawalmanindra @notsneakypotato https://t.co/x1wIDfkHH5

[A rose is @implisci]

@agrawalmanindra @notsneakypotato

[Flavio Brasil @fwbrasil]

Fable is a great model for meta-tasks. I couldn't see a major improvement in regular tasks but it's much better at introspecting and optimizing its own work. Paired with the support for nested subagents in CC, it feels a step change! 🤩

🧩A quite effective agent topology I've been using today:

✔️ Fable as the supervisor, handling high-value tasks like planning, self improvement, and strategic reviews
✔️ Opus as the coordinator, launching agents, validating the work, handling agent misbehavior, recovering from failures
✔️ Sonnet as the implementor, working on provided specs and plans, running tests, and reporting back
✔️ Haiku as the linter, matching changes against a database of anti-patterns under higher parallelism

All with skills and nested subagents. One layer launches the other 🥞

I still don't think this is AGI but there's no turning back for software engineering. The future is AI-first only! ✨

[A rose is @implisci]

@xu_si95720 @CryptoCyberia

[A rose is @implisci]

@xu_si95720 @CryptoCyberia

[A rose is @implisci]

@bryan_johnson Good chance the US consulate officer sees this thread. Good luck.

[A rose is @implisci]

@mcuban @costplusdrugs It’s useful for endothelial arterial function as well. The low dose 2.5 mg. I’ve found costplusdrugs to be very helpful. Low cost - no insurance needed.

[A rose is @implisci]

@kaavyya Took a 10500 mile drive across the USA and camped outdoors with a friend. Learnt driving on stick shift in the process. Had no idea of car breakdown or its consequences (did happen once in Colorado). When you are young you take risks and learn. You literally have nothing to lose.

[A rose is @implisci]

@SuperSisi Grok knows:

[A rose is @implisci]

Chebyshev only sets the upper bound for values k standard deviations away. Specific distributions have bounds that are much lower. It should be fairly simple for them (they have the data) to see what type of anomaly this is: is it 4 sigma (1 in 16000 for two tailed gaussian) or is it 6 sigma signal (1 in 500 million). My point is that just the addition of a couple of sigmas drastically lowers the probability. Now the distribution may be a heavy tailed gaussian but one has to be careful here. Whence the heavy tail. And so on.

[A rose is @implisci]

@afshineemrani How do you measure blocks in the cath lab. Isn’t angio the gold standard

[A rose is @implisci]

@shrutammegopaya Make a three strikes in an over and you’re out to make getting beaten dramatic.