Fawz

Plastic & Reconstructive Surgery Registrar | PhD Candidate Medical Data Law & Regulation @UoELawSchool | Sports, Policy, & Surgery | LLM / MBA

[Fawz @FawzNK]

Rare tweet from me nowadays. From a data law perspective, I think MedTwitter has misunderstood the new guidance. The law hasn’t changed. You access patients notes on special provisions for their health - not to check complaints (esp if not caring for the patient anymore).

[Conn @ConnCFC]

Merino when you need a goal off the bench

[Fawz @FawzNK]

@ajd90 I'm glad someone else finds it useful!

[Fawz @FawzNK]

Made a free-to-use website for doctors, nurses, and theatre staff to learn instruments used in theatre. Have a play here: https://t.co/CAbzrWVLgu

[James Steen @BMA_James_Steen]

Schrödinger’s doctors:

🔹Not important enough to restore their pay
🔹Not important enough to employ enough of them as 7m sit on waiting lists
🔹Not important enough to train enough become our GPs and consultants of the future

Whilst simultaneously;

🔸Too important to strike

[Kieran @kieranfbUK]

I was shocked to find a lot of NHS employers purposely getting rid of doctors just before they reached 2 years to avoid giving them proper employment rights. If it was in any other industry we’d happily call it what it is - exploitative

[Fawz @FawzNK]

Absolute shambles. I have said it time and time again, kangaroo court in action.

Failing system where doctors hold ultimately responsibility for non-clinical failures of the system.

I fear that this will be the same with software implementation.

[Mike Cosgrove @mikecosgrove]

Corporatism and contract rigging are killing the NHS.

A friend manages a department in a London hospital. Ordered 4 plug sockets to be changed, the firm took ages and charged £2k.

The department isn’t allowed to use any other company than the one assigned by the higher ups.

[BladeoftheSun @BladeoftheS]

Doctors are going to go on Strike again.

Nearly 50% of Doctors who passed their Medical Degree in the last 3 years are unemployed in the UK.

It used to be 1-2%.

Yet Waiting Lists are at record highs and not going down.

Why aren't they employed?

[Fawz @FawzNK]

There's so many doctors masquerading GPT wrappers as start-ups, that it's now getting a bit terrifying.

[Fawz @FawzNK]

Doctors will have a heart attack when they realise how medical software liability is apportioned.

[NIK @ns123abc]

>be OpenAI
>train Sora 2 on every movie/show ever made without permission
>release product with aggressive marketing

>Day 1-3
>Pikachu robbing CVS
>Hitler SpongeBob
>Mario, family guy, south park everywhere
>#1 app store in 48 hours

>Employee: "ser, we won"
>"users doing exactly what we planned them to do"
>new funding secured
>$500B valuation

>suddenly lawsuits incoming
>disney, NYT, everyone mobilizing

>Altman: "quick, make it the users' fault"
>"and copyright holders' fault too"

>emergency update ships overnight
>*media upload agreement popup*
>YOU agree not to upload copyrighted material
>YOU agree you have all rights to uploads
>misuse = permanent banned from sora AND chatgpt

>Liability successfully transferred:
>✓ users
>✓ copyright holders must opt out each character individually
>✗ openai

>Altman drops blog: "please use responsibly :)"

>users try to delete sora account
>deletes chatgpt too
>banned from ever signing up again
>API access? gone
>all data? deleted
>can't even use same email/phone
>users realize they're trapped

LOL, LMAO even.

[Rohan Paul @rohanpaul_ai]

🚨 BAD news for Medical AI models.

MASSIVE revelations from this @Microsoft paper.

🤯 Current medical AI models may look good on standard medical benchmarks but those scores do not mean the models can handle real medical reasoning.

The key point is that many models pass tests by exploiting patterns in the data, not by actually combining medical text with images in a reliable way.

The key findings are that models overuse shortcuts, break under small changes, and produce unfaithful reasoning.

This makes the medical AI model's benchmark results misleading if someone assumes a high score means the model is ready for real medical use.

---

The specific key findings from this paper 👇

- Models keep strong accuracy even when images are removed, even on questions that require vision, which signals shortcut use over real understanding.

- Scores stay above the 20% guess rate without images, so text patterns alone often drive the answers.

- Shuffling answer order changes predictions a lot, which exposes position and format bias rather than robust reasoning.

- Replacing a distractor with “Unknown” does not stop many models from guessing, instead of abstaining when evidence is missing.

- Swapping in a lookalike image that matches a wrong option makes accuracy collapse, which shows vision is not integrated with text.

- Chain of thought often sounds confident while citing features that are not present, which means the explanations are unfaithful.

- Audits reveal 3 failure modes, incorrect logic with correct answers, hallucinated perception, and visual reasoning with faulty grounding.

- Gains on popular visual question answering do not transfer to report generation, which is closer to real clinical work.

- Clinician reviews show benchmarks measure very different skills, so a single leaderboard number misleads on readiness.

- Once shortcut strategies are disrupted, true comprehension is far weaker than the headline scores suggest.

- Most models refuse to abstain without the image, which is unsafe behavior for medical use.

- The authors push for a robustness score and explicit reasoning audits, which signals current evaluations are not enough.

🧵 Read on 👇

[Fawz @FawzNK]

What a complete, utter failure of workforce planning, campaigning, & our union on every level. This is entirely predictable & manufactured. I remember yrs ago reading tweets like “don’t assume malice where there’s stupidity”. This is malice & stupidity.

https://t.co/ARvl1oTZyS

[Nicola James @NAJ562]

Had a CT scan this morning. On time and kind staff - thank you.

But it wasn’t the NHS. It was a private mobile unit parked at the hospital and paid for by our NHS.

This is PFI all over again: the public pays more so private firms profit.

The money:
▪️2018/19: £123.8m for 3.9m outsourced scans = £32/scan
▪️2022/23: £368.9m for 3.7m scans = £100/scan
That’s triple the price per scan and for fewer scans. Waiting lists are still high. So how is this helping us?

The staffing scam: same NHS trained radiographers/radiologists hired by private firms and sold back to us at higher rates. The NHS trains, and private companies cream off the profit.

The “efficiency” myth: I never received the first appointment letter, the next arrived at short notice with no text/email reminder. If a patient misses an apt, the trust still pays for the block booked session. Public money wasted - the private firm still gets paid.

The revolving door: We fund the NHS through our National Insurance payments. Politicians hand our money to private providers, those providers bankroll the politics that keeps it going. Everyone wins - except the public.

Our political system is broken. We need a new type of politics - policies that benefit ordinary people, decided by the people, and with representatives accountable to the grassroots members.

That’s why I’ve signed up to Your Party because I want the kind of policies and integrity that Jeremy Corbyn espouses - for the benefit of the many, not the few.

👉 https://t.co/cQLLw148Ya

#StopOutsourcing #EndPFI

[Fawz @FawzNK]

You all want AI documentation in healthcare until it documents something dumb and you sign off on it accidentally then the hospital, software developer, and solicitors say: "oh, we know we advertised it as near infallible, but you should have checked the output."

[Fawz @FawzNK]

So many great speakers now announced for the @PLASTAUK Big Ideas Event including:
@PlasticsFella, @Ryan_L_K, @Claribone_uk and speakers from @GoogleHealth !

Still looking for prize submissions to win one of 5 national cash prizes!

Find out more here:
https://t.co/lth3MzlQ4x

[Fawz @FawzNK]

On that note, on-call supplement of ~8%, weekend supp. of ~10%, and night supp. of ~37% are far below market rate. We can increase our salary by pushing for rates closer to 50% - in line with other jobs, much less public outcry & increase average salary well above FPR.

[Fawz @FawzNK]

Genuine question: if we want to get towards FPR w/o fighting the government on the PR battle, why does the BMA not start by changing social hours to something more social eg M-F 8.30-5? Will lead to fair pay rise for many (not all) doctors while also being difficult to criticise.

[Fawz @FawzNK]

@drumairrk True but it will help increase pay for medics doing on-call shifts, weekend shifts, long ward cover, as well as A&E staff on twilight etc; and, secondly, it is only a reversal to the 2016 changes rather than a “pay rise”.

[Fawz @FawzNK]

Genuine question: if we want to get towards FPR w/o fighting the government on the PR battle, why does the BMA not start by changing social hours to something more social eg M-F 8.30-5? Will lead to fair pay rise for many (not all) doctors while also being difficult to criticise.

[Fawz @FawzNK]

Welcome to the PLASTA Big Ideas Event!

🚀 Talks from surgeon-entrepreneurs & tech experts, a live Hackathon, & £1000+ in prizes! 🏆

Join us @ UCL East for a day of innovation & NHS future-proofing.

🎟 Tickets: https://t.co/znawW4Oay8
📝 Submit: https://t.co/vCs3ynklXK