Vish

MedComms & Pharma Marketing Leader | AI Strategy, Governance & Workflow Transformation | MD

Mody & Associates, Kepler26, KOL Luminary, Easy Annotation

[Vish @VishalsProjects]

What kind of KOLs or DOLs are you looking for? Just ask our system. It's super easy and the results speak for themselves.

Don't waste your resources on stale databases. Our fully agentic system finds everything. Experts that match your brief with current, verified, and source- linked information. Export your shortlist as beautifully formatted client and presentation-ready files: ppt, pdf, or excel.

All the leg-work done. 100s of hours and 100s of thousands of $$ saved.

We are aiming to be the best expert identification platform on market.

Founding Member pricing closes 31 December 2026 · limited to the first 25 members.

https://t.co/qF69uGHgex

#KOL #DOL #medcomms #medaffairs #msl #kolidentification #kolmapping

[Sebastian Caliri @SebastianCaliri]

Surgery cannot be mastered by simply reading books. It would be convenient if true, because training great surgeons requires decades of hands-on experience and we don't have enough of them.

What if there was another way to scale surgical intelligence? Mosaic Intelligence Labs was founded to solve this difficult and important problem. Today we are excited to introduce Mosaic to the world.

Mosaic's focus is on endovascular procedures - surgeries that use flexible catheters to reach deep inside the body through blood vessels, entering through a puncture just millimeters wide (often the femoral artery near the groin - see animation).

Mosaic is building sensing-enabled surgical instruments, spatial intelligence, and robotics. The immediate goal is to give surgical machines their own perception of the anatomy they navigate and expand what surgeons can perceive. Machine perception is a foundation for a future of reasoning and robotic actuation that assists surgeons.

If successful, Mosaic will enable more patients to access a broader range of interventions with less cost. It will give every surgeon the intuition and dexterity of the very best practicioners in the world, enabling new procedures and substituting invasive surgeries with access through blood vessels.

It is a mission that we are extremely proud to support at @8vc as the lead investor in Mosaic's $47M seed round.

Cofounders, the excellent @ilepirox and @alimackanic will be sharing more through the week.

And we are hiring in San Francisco! Read more at https://t.co/INkXFzUgk8

[bebu @bebuu_ai]

Wild rumor coming out of China this morning on DeepSeek V5:

DeepSeek is reportedly already pretraining V5 as a ~2 Trillion parameter ultra-sparse MoE model.

And instead of training on NVIDIA GPUs, they are reportedly running the V5 pretraining run on Huawei's new Ascend 950 / 950DT chips with a team of Huawei kernel engineers physically working inside DeepSeek's offices to keep the cluster humming.

If DeepSeek pulls off a 2-trillion-parameter frontier run natively on Ascend 950 silicon, that changes the entire AI hardware story.

[Vish @VishalsProjects]

I'm comparing KOL Luminary to new models being released to see how they compare.

Asked Google Gemini 4 Carbon to find a list of :

"Find 20 US oncology DOLs with expertise in ADCs for solid tumors for an advisory board."

It gave me a list, but no links or sources, no publications or clinical trial activity, no details at all.

I had to prompt it to find additional details and to also create an excel sheet for me.

It just took the list it had already come up with, added a few more details but not much more; still no links; nothing verified.

And then it told me to cut and paste it's table inside an excel sheet myself!! Even regular google search is able to give you an excel sheet.

CONCLUSION: Gemini 4 Carbon cannot compare to KOL Luminary for the purposes of KOL/DOL identification projects in pharma.

Gemini 4 is a capable model and the names it did find were good. But that means you still have to do all the leg work yourself for finding their relevant pubs, speaking engagements, industry relationships, network connections, etc.

[Vish @VishalsProjects]

@TheTimeNomad Tests being the product is true, but someone still has to decide what counts as correct. That part didn't get automated away.

[Vish @VishalsProjects]

@Souradip3000 Docs written for agents is the funniest plot twist. Humans never read them either, so the bots just got the first real audience.

[Vish @VishalsProjects]

@EAbrahams69 Six. Four people editing one file with an AI cheering them on sounds fun for ten minutes and chaos after that.

[Vish @VishalsProjects]

@Hasannshah Best part is it still works when your query is badly worded. Which is every query I write.

[Vish @VishalsProjects]

@tonyaibuilds Reels beating carousels at 19 followers is almost expected, the feed pushes video hard. Would test again once you hit a few hundred.

[Paras Chopra @paraschopra]

I asked Claude to make a novel discovery in any field it chose.

The only rules: it must be new, and checkable with high confidence without lab work.

It found evidence against a widely cited claim in cell biology: that mitochondria run 10-15 °C hotter than the rest of the cell.

[Vish @VishalsProjects]

@w_milczynska Add https://t.co/i3WBDdhbaF to b your list.

[Theo - t3.gg @theo]

The $200 Claude Ultra plan gets you $15,000 of Opus 5.5 in a month, but only around $3,000 of Fable 5.5.

Even ignoring that Fable is restricted to 50% of your weekly limit, it burns faster per $ than Opus does.

More info below

[Claude @claudeai]

A thread of early experiments with Claude Fable 5.2:

A 3D model of a cell, built entirely from code one organelle at a time. Built by @kevin_t_ngo.

[Vish @VishalsProjects]

@DeryaTR_ @bot What is the total $ cost for this video? Estimate.

[Vish @VishalsProjects]

@DeryaTR_ @ElevenLabs @higgsfield What was the total $ cost for this video?

[Vish @VishalsProjects]

@DrSusanNasif There's nothing simple about agentic solutions that bring down hundreds of hours of manual work down to minutes.

What exactly is "capacity building"?

[Vish @VishalsProjects]

@sridharfyi Niche, pharma industry product
https://t.co/8vQb95qB4B

[International Cyber Digest @IntCyberDigest]

‼️BREAKING: Grok Build has left 35,000+ of its users' apps open to anyone, including apps from SpaceXSI.

The open apps held 3,784 medical exams of workers and a Tesla technician scorecard, plus a Navy contractor's roster and SpaceXSI's own sales plans.

The "secret" words in each app's link sit in its web address, which DNS logs record, security firm Claroc found.

Cursor, which SpaceXSI bought, says apps are private by default. Claroc's tests show otherwise.

SpaceXSI pulled its 24 apps after Ars asked. Most other data was still exposed when Ars published.

[Vish @VishalsProjects]

@sujal_twt Simulating ideas before betting on them is a nice angle. Curious how you'll keep the sim from just agreeing with you.

[Vish @VishalsProjects]

@swandexapps Good point. Everyone's rushing to redraw roles around AI and forgetting the people who actually know what the regulator will push back on.