Navin Ramachandran
navinram.bsky.social
Navin Ramachandran
@navinram.bsky.social
Always building. Visiting Professor Dyson School of Design Engineering, Imperial College London. Radiology Doctor UCLH. Supervisory Board IOTA Foundation.
Tbh I love London as a centre of working but the major issues are red tape and fund raising. Much easier to raise in America with much larger cheque sizes. And the regulation is often a bit more lax (not good for fast innovation, but may be better for the world in some ways 😅)
May 21, 2025 at 5:01 PM
The whole registry / research landscape is being challenged with the recent NIH changes.

It is a difficult time for many patient groups and researchers.

Can we build something that rises to that challenge? We hope we can and we would love to speak to anyone who shares the same vision.
April 23, 2025 at 10:56 PM
So the CEO of a company selling AI agents suggests that we will replace our staff with AI agents?
April 22, 2025 at 2:51 PM
Correct!
April 2, 2025 at 10:33 PM
Sorry to butt in Veni. I truly empathise with dealing with a suffering child.

I am no expert on corticosteroids but prolonged use can be problematic. And can lead to steroid withdrawal reactions which may be what is happening.

bnf.nice.org.uk/drugs/fluoci...

Could that be the case?
BNF is only available in the UK
The NICE British National Formulary (BNF) site is only available to users in the UK (England, Scotland, Wales and Northern Ireland)
bnf.nice.org.uk
March 10, 2025 at 9:07 AM
“would need access to”

🙏edit button🙏
March 5, 2025 at 9:23 PM
I see this runs on a particular mobile hotspot as a PoC.

Sorry to ask a stupid question but could this Rust code be reused to make mobile apps for iOS / Android which does the same? Or are there restrictions on the device functions it would access to on the mobile platforms?
March 5, 2025 at 8:32 PM
I was listening to @edzitron.com talking about why he thinks a grok-like generative AI is what is missing from this platform.

I think the 2 of you should have a chat…
February 9, 2025 at 7:05 PM
That’s good enough for me!
February 9, 2025 at 7:00 PM
Never been a toastie lover but if someone loves it as much as carnitas then it’s worth a look when I am next in the area 😂🙏
February 9, 2025 at 6:58 PM
Home
www.kappacasein.com
February 9, 2025 at 6:47 PM
Curriculum for modern life
a cartoon of a chef with the words perfect written above him
ALT: a cartoon of a chef with the words perfect written above him
media.tenor.com
February 5, 2025 at 11:00 PM
Random @mcuban.bsky.social ping.

Amongst the millions of messages you get. 😄

But I would love to understand why you prefer the subsidy over the cap.
January 29, 2025 at 11:08 PM
There may be more far reaching consequences of these approaches and I would really appreciate if someone could outline the long term benefit of the subsidy vs cap 🙏
January 29, 2025 at 11:04 PM
I am not sure I agree with this.

A cap limits the price and the amount of profit.

A subsidy does not limit the price or profit but needs to be paid from the public purse.

So from a very basic assessment I would prefer the former.
January 29, 2025 at 11:02 PM
Any chance of verifiable credentials using the atproto DID?
January 28, 2025 at 6:12 AM
In the non-Sam Altman / MS contract take:

A push to replace the majority of jobs with a less capable but
replicable and 24/7-available workforce.

They don’t have to be good at a job if they are quick and always available.

TLDR: bad times.
January 17, 2025 at 12:18 AM
Do you still have any Meta products installed?
January 15, 2025 at 5:39 AM
Very good points Kim. But I am not sure who is leading in AI atm

www.chinatalk.media/p/deepseek-c...
Deepseek: The Quiet Giant Leading China’s AI Race
Annotated translation of its CEO's deepest interview
www.chinatalk.media
January 14, 2025 at 5:21 PM
Can be trained - sure.
What has happened in reality - not so sure.

techcrunch.com/2025/01/08/e...

Apologies but my point was more generic about things existing but people being allowed not to like it 😅
Elon Musk agrees that we've exhausted AI training data | TechCrunch
Elon Musk concurs with other AI experts that there's little real-world data left to train AI models on.
techcrunch.com
January 13, 2025 at 1:48 PM