Lawrence @Deep_Chora

MLE with a focus on Computer Vision. Long COVID since September 2022. I allow myself to change my views, because ideas are meant to be tested, not defended. Joined August 2025

Tweets

228
Followers

47
Following

512
Likes

2K

Lawrence @Deep_Chora

6 hours ago

@peterottsjo I see, then we are in agreement. I specifically rail against the LLM hype, however, because the problem structure requires that we train on actual biological data, not just language, and we will fail if we just rely on that.

0 0 1 16 0

View Details

Lawrence @Deep_Chora

15 hours ago

@peterottsjo The first two are classical Deep Learning models, not LLMs. The third one just uses LLMs for choosing tools, fourth one is again classical deep learning, etc. You are kind of making my point for me, which is that classical approaches are superior to LLMs in biology.

1 0 0 48 0

View Details

Lawrence @Deep_Chora

2 days ago

@kelseyshields08 Hey Kelsey, do you mind sharing what treatment approaches he discussed? My experience is that Microclots are very non-specific. They indicate some type of inflammation, but it is hard to say which kind.

1 1 5 627 1

View Details

Lawrence @Deep_Chora

2 days ago

@JackHadfield14 When do you reckon you'll share what you found with the scientific community? Are there any plans to publish a paper?

1 0 0 49 0

View Details

Lawrence @Deep_Chora

2 days ago

@Will52773T @dantawfik @HermanPontzer Presumably this work was done on a very constrained and limited type of exercise. Don't even need to read the paper for that.

0 0 1 320 0

View Details

Bill Gurley @bgurley

3 days ago

If you are on the verge of AGI or ASI, why isn’t your model smart enough to recognize espionage distillation in real time? You say “cure cancer in a few years.” Isn’t sniffing illicit distillation quite a bit easier than curing cancer? Why write letters to DC? Just use AGI.

277 456 6K 500K 614

View Details

Andrew Hammel @AndrewHammel1

3 days ago

The cardiac unit of the Düsseldorf University Hospital, where people recover from heart surgery, is currently measuring 38°C/ 100.4°F. As you can see, the building is only about 15 years old. And built without central air-conditioning.

542 1K 7K 1.2M 669

View Details

Lawrence @Deep_Chora

3 days ago

@LocasaleLab Another point, how are you gonna use RL in a system you cannot cheaply simulate? Sounds a little naive.

0 0 0 53 0

View Details

Lawrence @Deep_Chora

3 days ago

@LocasaleLab Using AI to cure disease would require vast amounts of bio-data. No tech company is willing to put up the money, though, because promises about "curing cancer" were always about justifying ever more absurd scaling efforts to concentrate power, not raising QOL or advancing cures.

0 0 0 252 0

View Details

Lawrence @Deep_Chora

3 days ago

@JackHadfield14 Tech != Biotech. 2-3 years is way too optimistic. Nothing happens unless you do it yourself in this space.

0 0 0 18 0

View Details

Mistral AI @MistralAI

6 days ago

Introducing Mistral OCR 4. It creates structure with bounding boxes, block classification, and inline confidence scores in 170 languages. 🧵👇

206 482 6K 1.3M 2K

View Details

Lawrence @Deep_Chora

a week ago

@Knesix That's how you think when you haven't seen research fail you again and again. That's how you think when your quality of life doesn't depend on a new treatment being found.

0 0 0 12 0

View Details

Jeffrey Emanuel @doodlestein

a week ago

When you see the doctors coming out in hordes to confidently argue against new scanning technologies powered by advanced math and statistics, just remember that the vast majority of them are borderline innumerate and can’t even understand what p-values are:

91 42 397 98K 64

View Details

Perry E. Metzger @perrymetzger

a week ago

I’ve seen this trope posted by a lot of doctors in the last 24 hours. The claim from doctors is that Midjourney’s cheap full body scanner is bad, because people would get treated unnecessarily for harmless conditions, and that would increase death rates. Now, I have a lot of skepticism about Midjourney’s announcement (a lot of skepticism) but I think that the doctors are failing to look in the mirror. What they are saying is, in effect, that they cannot be trusted with having more information about patients. That, if you give doctors too much information about a patient’s condition, they may kill the patient by treating them unnecessarily. This isn’t a problem with inexpensive full body scanners. This is a problem with doctors and medical practices.

Jeremy S. Pollock @bmorecardiology

a week ago

Same issue w full body MRI.. indiscriminate screening of healthy people causes harm Making the scan quicker and cheaper… likely will increase the scale of the harm. @CanesDavid @afshineemrani @grok explain harms in screening a population w low pretest probability for disease

101 4 95 326K 31

185 119 2K 166K 182

View Details

Matt Quinn @Mqsley

2 weeks ago

No idea if Midjourney is selling vaporware, but the takes from doctors are wholly embarrassing for an industry that fails most people in some way

68 32 667 39K 28

View Details

Lawrence @Deep_Chora

a week ago

@D__Melb @Geero_Gero @gfodor That might have been true in the past but is no longer true. The low-hanging fruits are gone. Novel research insights are going to be disproportionately generated by methods underused in a field, one of which is big data biomarkers. It's not about vanity, it's structural.

0 0 1 40 0

View Details

Prof. Karl Lauterbach @Karl_Lauterbach

2 weeks ago

29% sind nicht Deutschland. Die Mehrheit der Deutschen will nichts mit Ihnen zu tun haben. Ein Weimar kommt nicht wieder. Wenn die Wirtschaft sich erholt, wir durchlaufen derzeit schwere internationale Krisen, geht es wieder bergab mit Ihnen. Regiert haben Sie dann nie…

Alice Weidel @Alice_Weidel

2 weeks ago

Die Union fällt auf 20%, unsere AfD hat mit 29% inzwischen 9 Prozentpunkte Vorsprung. Deutschland will den politischen Wandel - wir stehen bereit!

1K 5K 32K 1.1M 373

4K 165 2K 494K 164

View Details

Lawrence @Deep_Chora

2 weeks ago

@NBoydGibbins @amaticahealth I'd be more impressed if you could separate patients by disease burden (mild / moderate / severe) based on their markers. The aim should be that a simple logistic regression can achieve this. Supervised classification is the only way.

0 0 0 28 0

View Details

Lawrence @Deep_Chora

2 weeks ago

@NBoydGibbins @amaticahealth Feeling tired after a marathon and experiencing severe PEM after taking a shower are not the same phenomenon differing only in magnitude. To assert so would be ridiculous.