Amplify Dude

@parmita We just built this. You’ll see this pushing live over the next week or so

Target ↓ Molecule ↓ Cell response the only form of discovery Isomorphic, Atomwise, and any generative chemistry companies allow for. Cell response ↓ Pathway ↓ Target is completely missing at scale. morphology does not inform pathways alone unless your cell response is categorical. anything that is on a continuum requires a better measurement of *phenotype* including cell chemistry. scRNA-seq predicts transcriptional activity, not the actual phenotype. we do not understand genes well enough to ever make that assumption. we can recover some new things, but it is a tiny fraction of what's actually happening in the cell. all forms of *AI* drug discovery today rely on assumptions like: - we already understand all pathways and can build a predictive response for a drug - we already understand what genes do well enough to map transcription onto new pathways - we already understand the target across different organs - we already know that binding is not cell-type specific and tissue-specific, and any differences can be recovered by cell lines. these are huge assumptions. ridiculous assumptions. observability is the new frontier, without which we cannot even BEGIN the conversation of completely new discovery.

@danieldwalton Spoke to Milo yesterday. Give him a raise haha

@amplifydude AI is one of the best nothing does it better than AI

AI - you’re either too lazy to understand it or becoming superhuman from it. AI is one of the things you either embrace and go deep into understanding to be able to use efficiently, or you half use and end up thinking it’s not an efficient spend of capital. We saw this with Microsoft recently with them racking up a $500M bill and deciding to cut AI spend from the business. This is where startups and solopreneurs can now compete in a field with the big guys and out perform them provided they have a breadth of understanding that spans across different business verticals to make better decisions. If you think that AI is inefficient, know that there are thousands of other people working harder than you to understand it and are succeeding tremendously with it in most cases at least increasing their output by x25-x100. Start understanding how agents interact, think, perform and output. Understand their weaknesses and strengths, how different models compliment one another and how you can slot this into solutions that improve your overall output.

@paidcircle I know a couple people

who can unban an IG

We have just built the layer you add on top of any model to get near perfect results across all fields. We’re currently beating a bunch of bench marks. The best thing is that it’s BYO model which then calls our engine via API and gives you insane ability. We’re about to release this publicly. We expect a lot of demand around this and have a few clinics already agreeing to trials waiting to get their hands on it.

For two years I have heard that we need specialized, medically-tuned models for clinical decision support. That frontier models are not safe or accurate enough for real medicine. Most of the people making this claim had clearly not been using multiple frontier models when they said it. Nature Medicine just published the test. An independent NYU group pitted two specialized clinical tools, OpenEvidence and UpToDate Expert AI, against three frontier models from OpenAI, Google, and Anthropic. Three benchmarks. One hundred real physician queries. Twelve blinded clinicians scoring the outputs. Frontier models won all three. Gemini and GPT led on medical knowledge. All three frontier models clustered at the top on blinded clinician review. And here is the number that should end the argument: the specialized clinical tools scored about the same as Google Search AI Overview. The purpose-built medical AI matched a search engine. To be fair, the authors are careful and they should be. Deeply subspecialized tasks may still favor domain-specific models. Hospital-specific systems grounded in institutional data are likely the future. This is a snapshot, not a permanent verdict. But the claim that you need a specialized model to get good medical reasoning from AI is done. Blinded clinicians killed it. One thing I will add from my own workflow. Frontier models still hallucinate a percentage of studies and citations. When I need to confirm a specific trial exists and says what a model claims, I check it on a retrieval-grounded tool like OpenEvidence. Not because it reasons better. It does not. But verifying a citation is a retrieval problem, not a reasoning problem, and those belong on different tools. The capability is here. The real skill is knowing how to use multiple models together, each for what it actually does best, with the clinician making the call on which is which.

picture this 8 of the world's biggest problems getting solved in the next 12-18 months 1. obesity: retatrutide phase 3 confirmed 30% bodyweight lost. 65% of patients no longer clinically obese. FDA submission late 2026 2. testosterone decline: FDA is expanding testosterone therapy. peptide and endocrine protocols going mainstream. the 1970 baseline is coming back. 3. birth rate crisis: embryo optimization commercially available today. scientists just rejuvenated aging human eggs in the lab. fertility is soon no longer a countdown 4. aging: Life Biosciences just injected the first reverse-aging drug into a human. Sinclair's oral reprogramming pill entering XPRIZE trials. 5. Alzheimer's: Retro Biosciences dosed the first humans with a pill that reactivates the brain's cellular cleanup machinery. Phase 1 results Q3 2026. 6. cancer: daraxonrasib nearly doubled survival in pancreatic cancer. RAS has been undruggable for 40 years. they drugged it. 7. mental health: psychedelics got a presidential executive order. Compass weeks from the first FDA approval of psilocybin. 8. heart disease: inflammation replacing cholesterol as the primary target. the root cause is finally being treated not the symptom. every single one of these has a clinical trial or an FDA action behind it right now humanity is slowly healing bio/acc

Some of you said it still looks like AI slop, and I agreed, so I told Fable to "100x the design." Here's what it did, and all the details on how.

Claude Fable 5 + Claude Design is f*cking insane 🤯 Anthropic just dropped its most intelligent model ever, and the first thing I pointed it at was email design. I built a complete email campaign design in Claude Design, and the difference is night and day: tighter layouts, cleaner hierarchy, on-brand from the first generation. All inside Claude Design with Fable 5. Perfect for DTC brands and agencies who are still paying email agencies $3-5K/month for campaign designs that take 2 weeks to ship. If your campaign calendar is packed but every new email means briefing a designer, waiting on mockups, sending notes, and waiting again... This workflow eliminates the entire bottleneck: → Load your brand design system into Claude Design once (colors, fonts, logo, button styling) → Switch the model to Claude Fable 5 — Anthropic's new state-of-the-art model with the best vision of any AI → Prompt the campaign email section by section: header, hero, headline, offer block, CTA → Fable 5 nails layout and brand details that older models fumbled → Iterate inline — swap images, adjust styling, color-pick directly in the canvas → Export the finished email and hand off to your ESP No briefing a designer. No 2-week turnaround on a single campaign. No paying an agency $4K/month for 4 emails. What you get: → Campaign emails designed in minutes, not weeks → A reusable design system every new email pulls from automatically → Noticeably smarter design decisions from Fable 5's upgraded vision → Full inline editing before anything touches your ESP Built 100% with Claude Design + Claude Fable 5. I recorded a full walkthrough showing exactly how this works. Want it for free? > Like this post > Comment "FABLE" And I'll send it over (must be following so I can DM)

@gnosispay @benjarosetree Hey legends, I just spoke to one of your team and my funds have been restored. Great work on getting everything going again.

Ben said it cleaner than we would. Account is a Safe. SEPA straight in. EUR stays EUR. Cashback comes out of program structure, not inflation. Still room to improve, for now. Great to have you back, @benjarosetree.

@0xchim9 @gnosispay @benjarosetree I’m in the chat logged in with support. It’s THEE gnosis support. I’ll checkout discord too

@amplifydude @gnosispay @benjarosetree Are you sure that you talked to the right support? Maybe share the ticket number, I'll forward to the support team... Otherwise you can come on the Discord and we can check this together... discord.gg/gnosispay

@0xchim9 @gnosispay Ok cheers

@amplifydude @gnosispay Yeah sorry about that, a few accounts will be back tomorrow/Monday

Card services are back to normal for over 99% of users. To protect your account security, every Gnosis Pay Safe has been replaced. The new Safes are integrated with your existing physical and virtual cards. Please note: your Safe address has changed. ONLY deposit to the Safe currently that is now shown in your account. Anything sent to the old card address (on chain or via IBAN) will be lost.

You get paid for the speed between idea and execution. Being fast and smart is a 1-in-a-million rare skill.

@koeppelmann @gnosispay Great to hear

For @gnosispay users: You will be able to continue using your physical and virtual card(s), but they will be linked to a new onchain Gnosis Pay Safe. Make sure to NOT send funds to old one anymore as they will be likely lost! All old Pay Safes were vulnerable and have therefore been replaced. Depending on the owner structure, only some were actually compromised and drained (by either white or black hat). If your old Safe was drained, your new Safe will be funded by Gnosis. If it was not compromised, you will need to migrate your funds to the new Safe.

@LoicDelobel This is such a great take

Most DTC operators dismiss telehealth because "it's healthcare". What it actually is: ecommerce with one extra step (a clinician review) and three extra benefits (recurring revenue, regulatory moat, real customer outcomes). Same skill set. Different ceiling.

NEW: Scientists have identified blood proteins that can reportedly predict lung cancer more than five years before diagnosis.

@danieldwalton Thanks mate haha

@amplifydude you are crushing it you will connect with even more keep posting

I don’t know what Elon did to the algo but X is the new LinkedIn. I’ve connected with unlikely 5 people this week just from posting on X and also being shown the right post in my feed with such laser precision that it’s like the algo knows my thoughts.

It looks like we will all be on some dose of a GLP-1 sooner than later. More and more incredible non-weight related benefits are being discovered as clinical studies continue. - Cancer progression is crushed - Livers completely heal - Cartilage regenerates in joints - Powerful immune modulation ...and so much more...plus all of the benefits we've yet to identify.

@LoicDelobel Do you set telehealth businesses up? We’re looking for telehealth partners in the US

You can read a thousand telehealth threads or open one tab. sanare.me is the Tagada Rx stack live in production: quiz → clinician → pharmacy → checkout → upsell → lifecycle. If you want to inspect the engine before talking, that's the cleanest way.