Stefano Pacifico (@StefPac) profile picture

Founder, CEO @epistemic_ai - AI agents x Pharma strategy - #rhabdomyosarcoma dad.

New York
Joined August 2009
Photo by Stefano Pacifico (@StefPac)Photo by Stefano Pacifico (@StefPac)Photo by Stefano Pacifico (@StefPac)Photo by Stefano Pacifico (@StefPac)Photo by Stefano Pacifico (@StefPac)Photo by Stefano Pacifico (@StefPac)Photo by Stefano Pacifico (@StefPac)Photo by Stefano Pacifico (@StefPac)Photo by Stefano Pacifico (@StefPac)Photo by Stefano Pacifico (@StefPac)
how does slack know how old I am !?!
39
Stefano Pacifico retweeted
We gave GPT control of a real microscope during a 72-hour cancer drug screening experiment. It imaged 3 times as many drugs as standard approach and demonstrated self-improvement by rewriting its own imaging scheduler six times. Details below:
4
4
33
2,953
Stefano Pacifico retweeted
This whole thing sucks, and I wish none of it had happened. It’s a really sad day! But I want to make one thing clear: we’ll keep fighting for open-source AI, harder than ever. This is too important to slow down now. Let’s go!
447
940
161
10,568
614,325
Stefano Pacifico retweeted
*NVIDIA PREPARING TO OUST CEO JENSEN HUANG: WSJ
95
189
206
2,014
527,779
Stefano Pacifico retweeted
My AI agents found an unmapped Maya site in Mexico, using LiDAR data from a NASA dataset. 72 hours from a random thought on vacation to a reply from Mexico's national archaeology institute. What a time to be alive…
263
2,048
621
20,039
1,380,343
Stefano Pacifico retweeted
A cone or a funnel
3
7
1
29
2,135
"Io la lavatrice la presi e la pagai" versione 2026
Ho passato giorni a dialogare con ChatGPT per scegliere lavatrice e asciugatrice. Abbiamo confrontato modelli, misure, prezzi. Ho ordinato, pagato, aspettato. Poi gli elettrodomestici sono arrivati, ma non vanno bene. Ecco cosa mi ha risposto ChatGpt
42
Stefano Pacifico retweeted
The ring appears to rotate counterclockwise, with either the right side appearing closer and the left side farther away, or the left side appearing closer and the right side farther away.
4
7
4
61
6,221
Stefano Pacifico retweeted
Want your own Slack for $0? I open-sourced Ember. It deploys to @Cloudflare’s free tier in about a minute, works with Slack-compatible payloads, and includes an AI assistant and media attachments.
170
171
99
2,626
409,822
Say it with me: I shall use my own evals
Nice DeepSWE score! Does it survive a lowercase filename? Introducing MaxxBench: 89 DeepSWE tasks with small changes to the instructions or starting code. Same engineering problem, slightly different requirement. Models clustered around 74% on DeepSWE end up 23 pp apart. Good at coding, or good at DeepSWE? 🧵
2
56
Stefano Pacifico retweeted
The squares in this checkerboard pattern are aligned both vertically and horizontally, but they appear to form an undulating pattern.
18
83
13
796
42,950
Stefano Pacifico retweeted
Read the fine print. This claims a $1M Phase 2 trial. What it actually contains is a 24 person Phase 2a repurposing an already approved drug that only lasted 2 weeks. Critically underpowered. The submissions are a mess, with a huge delay of registration over 100 days after study start, parallel vs crossover discrepancies, an unfavorable finding from ethics board, and notably a failure to report the results in January as was committed. They benchmark against trials much larger than theirs to make their numbers seem lower in a misleading way. The study uses an existing oral medicine, one virtual site, home-based monitoring, and a very small patient population. These are all important cost-reduction strategies, but none depends uniquely on generative AI. I’m supportive of AI reducing trial costs but this is wildly misleading and unhelpful.
7
13
1
105
7,222
Stefano Pacifico retweeted
円盤は時計回りに回転して見えるが、要素のドットは行ったり来たりしているだけである。
2
7
2
88
5,177
Stefano Pacifico retweeted
Do some cool research, get $25k free Modal credits! Really looking forward to see what comes out of this.
Over the past few months, our team visited universities to meet students doing AI research. We left convinced that some of the most original thinking in the field is coming from undergrads. We met students studying how models reason, building new evaluation methods, and applying AI to science and medicine. Their ideas were ambitious and their research was rigorous. And nearly all of them were held back by the same thing: compute. The GC Research Fellowship, in partnership with @modal, gives 20 exceptional undergraduates up to $25,000 each in compute to pursue their biggest questions. Applications are open: generalcatalyst.com/research…
3
8
5
103
12,069
Stefano Pacifico retweeted
Having the interviewer steal $5.9B from your current employer is just a brutal way for your company to find out you’re interviewing for jobs
genuinely impressed by the level of sophistication of this hack. this is an entirely different league from those pathetic CVE exploits
24
116
5
3,112
140,923
If you work in pharma, pharma services or AI, you know it’s moving at a rapid pace, but also not fast enough. Here are my recent thoughts on where pharma services are headed, along with a few anecdotes from our work @epistemic_ai
1
42
Stefano Pacifico retweeted
🚨Last chance 🚨 10 Assistant/Associate Professorship openings at Hasso Plattner Institute (HPI) & University of Potsdam, Germany Topics: AI / computer science / societal issues of technology / digital health etc. HPI is a non-profit academic institute seeking to hire several dozens of new professors in the coming years. The positions are also affiliated with the public University of Potsdam. Apply until Oct 31: faculty-search.hpi.de/jobpos…
1
12
33
5,333
... and by being smarter about what and when to bring to the clinic (or to the next stage).
Despite staggering advances in basic science, drug development research is only getting slower. The main culprit? Increasing clinical trial timelines. The median clinical development pathway has increased from 8.4 years to 9.3 years and pipeline creation has slowed 32% even as R&D spend has gone up. This is the point I keep making about the main bottleneck to drug discovery and why AI on its own won't be enough. These timelines aren't going anywhere (ofc AI can be used creatively to ease such bottlenecks -- eg via better endpoints that allow smaller cohorts for trials, followed up for shorter periods of time).
1
44