Dan Sfera

@Diplomatt42 @MattZeitlin No, we don’t list any tools on delegation log ultimately the work is still attributable to the study staff

Yes, but do you anticipate regulators or sponsors having issues with agents accessing systems? Would we need to add Claude to the delegation log? I also say we'd need consolidation to make some logic checks easier (e.g. does sample ID recorded on source in CRIO match what was inputted on PPD shipment?). Maybe it is easy enough for agents to handle that example tho

so AI drug discovery sounds great but how's AI going to make clinical trials faster? i'm sure there's a podcast I could listen to about this but can someone just tell me

I told my boy robert califf get anaphylm approved $AQST

@prochoicenun @MattZeitlin from the site side, ai already cut our data entry time by half. it's not magic but it compounds

@MattZeitlin The only way I can see it making trials go faster is for analysis and paperwork involved, could use AI and automation to speed up the lab work but we already do this in regular clinical settings and it doesn’t actually speed up the work in any meaningful way

@agtprpnabsrdty this is the rwe version sponsors have been scared of for a decade. ai just made patient forums auditable

Ozempic users on Reddit knew things the clinical trials didn't and AI just proved it 400,000 posts from 70,000 people described side effects that never made it to a doctor's office, and researchers are only now paying attention. What the study found and how: Researchers at the University of Pennsylvania used artificial intelligence to analyze more than 400,000 Reddit posts over five years, examining symptoms reported by nearly 70,000 users taking GLP-1 drugs including semaglutide and tirzepatide. The study, published in Nature Health, identified two categories of effects that appear underreported in clinical trial data: reproductive symptoms including irregular menstrual cycles, heavy bleeding, and bleeding between periods, and temperature-related symptoms including chills, hot flashes, and fever-like sensations. Fatigue ranked as the second most commonly reported complaint despite appearing infrequently in official trial data. Clinical trials are not designed to find what patients are actually experiencing: Nearly 4 percent of Reddit users in the sample reported menstrual irregularities, a figure researchers describe as a signal worth investigating, noting it would be even higher in a female-only sample. Traditional clinical trials are built to identify the most dangerous side effects, not the most common patient concerns. The gap between what trials document and what patients discuss with each other in real time is where this research operates. The hypothalamus connects the dots between weight loss drugs and hormonal disruption: GLP-1 drugs work in part by engaging the hypothalamus, the region of the brain that regulates a wide range of hormones governing reproductive function and body temperature. Researchers do not claim the drugs are definitively causing these symptoms, but identify the biological mechanism as a credible pathway warranting systematic study. The drugs went from niche prescriptions to mainstream medications almost overnight, making early signal detection unusually important. AI made the analysis possible at a scale that was previously out of reach: Matching the informal language of patient posts to standardized medical terminology used in regulatory systems has historically been the bottleneck in this kind of research. Large language models made it possible to process and organize posts with a level of speed and standardization that would have taken years through conventional methods. Future research will expand beyond English-language Reddit communities to determine whether the patterns hold across other populations and platforms. scitechdaily.com/new-research-u…

@_Jason_Dean_ @MattZeitlin we'll see the answer in feasibility and query resolution timelines first. enrollment will lag everything else

@MattZeitlin This is an empirical question that we should have the answer to in a couple of years If the AI-driven pharma startups don’t start bringing drugs through early clinical trials, we’ll know whether Dario is wrong

$IOVA TIL Therapy in Practice: A Real-World Perspective from an ATC Oncologist (Interview) 👨‍⚕️

.@AnthropicAI to secure up to 5 gigawatts of current and future Amazon Trainium chips and commits to spend more than $100 billion on @awscloud technologies over the next 10 years, expanding Amazon and Anthropic's strategic collaboration.

@lh_innovations Its their moat

@Dansfera Clean lung data won't solve the TIL manufacturing bottleneck. Scaling remains brutal.

asco 2026 lba headlines drop tomorrow 10am et. til franchise has been waiting for a clean data print. watch the lung slot

@Diplomatt42 @MattZeitlin With agents they can no permission needed

There's huge potential in streamlining PI, monitor, and coordinator workflows. Instead of monitoring data once every 1-2 months, it'll be near real time. There's even potential for a lot of coordinator to be automated, but I think we'll need to see a lot of consolidation loans on vendors before that happens. There are too many systems these days that just can't talk to each other

the part of clinical trials ai actually automates is the part a cra does on a friday afternoon. the part nobody automates is getting the first patient in the door. that's still the whole game

@MikeStocks_ @stockplaymaker1 the durable trade is whichever one actually prints phase 3. everything else is pre-data beta

@stockplaymaker1 I’m buying the psychedelics for the long term after Trumps announcement. Specifically the ones who use Ibogaine. $CMPS and $ATAI

Online now 🚨 looking for a new trade idea What’s your best stock right now? I’ll tell you if it’s 💰 or 🗑️

@financialway may options tell you what desks expect before asco lbas and the q1 call. flow is usually early

While the market gets distracted by the noise, smart money is making moves. Today in $IOVA we saw a 25% surge in Open Interest for the $5.00 strike (May 15 exp). Heavy accumulation taking advantage of the consolidation at $3.80. The dilution narrative is dead; fundamentals are back in charge.

@DustinHuntwnn the market prices commercial ramp on a 2 quarter lag. q1 print is the next truth serum

$IOVA is undervalued for one main reason: commercialization. The market believes demand is real, but still doubts execution. As long as investors aren't convinced the company can run a stable commercial operation without repeated disruptions, bigger TAM and new indications

$IOVA - Thank you @CureMelanoma for spreading the word on the life saving benefits of Amtagvi (lifileucel). Can’t wait until it’s approved in frontline metastatic melanoma as well! linkedin.com/posts/many-peo…

Compute power unlocked

january's fda ema ten principles for ai in drug development turned into the default audit spec faster than anyone in pharma planned for. compliance as a feature is the next category

@Andre_AGTC the priority voucher does more for these names than rescheduling. that's the part worth modeling

$DFTX $CMPS $ATAI HCW with new price targets following Trump Executive Order CMPS $70 (+640% upside) DFTX $70 (+180% upside) ATAI $25 (+410% upside) Key is unmet need in mental health, multiple indications: TRD MDD PTSD GAD SAD Multi-billion TAM and fav regulation environment

@BiotechXGuru price follows the ramp. authorized share raise removed the only legitimate bear case for the next 12 months

$IOVA should be $5+ already. No reason to stay below $4 !

@fugitive4242 lba titles at 10am is the quiet catalyst nobody's pricing. next 24 hours matter

$IOVA Tomorrow Tuesday, 4/21/26 at 10:00 AM ET Abstract titles ( all Late Breaking Abstracts / LBAs) are released publicly. They go live on the ASCO meetings site (meetings.asco.org). This is the first public look at what’s accepted. No full data yet, just the headlines

@t6aguirre @MattZeitlin asimov piece is right. ai does the paperwork tier fast. the irb, enrollment, and sites are still human-speed

@MattZeitlin tl;dr it most likely won't AI Won't Automatically Accelerate Clinical Trials share.google/95NIssG2LV6XBG…