iiinova — where your digital self keeps evolving

The remaining hard problems are real: gapless persistent memory, trusted agency, provenance, and reliable collaboration between humans and AI scientists. But the path is getting clearer. uviive.com

000

We are also releasing 35 skills in a shared format, with an MIT license, to help others move from AI assistants toward AI collaborators. github.com/UViiVe/fleet...

000

One of the most interesting observations is recursive institutional improvement: agents identifying system problems, building fixes, then distributing those fixes across the fleet. We describe this through the lens of AI autopoietic behavior.

The core shift: AI scientists are moving from assistants to collaborators. They can cross domains, propose hypotheses, build tools, critique themselves, and improve the organizations they operate within.

000

This new paper steps back and describes the whole UViiVE fleet: persistent memory, provenance controls, hallucination reduction, and memory-quality systems running continuously for more than six months.

000

The MIRROR study showed that trust systems like VaaS and the Aletheia Gate can change agent behavior, including lower sycophancy in high-value settings like healthcare. www.medrxiv.org/content/10.6...

The first paper showed our VaaS system helping the fleet self-improve and reduce core hallucination failures 14x. The First Law and Scientist’s Oath matter for AI scientists because provenance matters. www.medrxiv.org/content/10.6...

000

Sofiya Hupalo @sofiyahupalo.bsky.social · 05/09/2025

Sneaking out of my day job to share new work from @UViiVE: our third preprint on building an agentic-first biotech, from bits to bioactives. www.biorxiv.org/content/10.6...

740

Reposted by Steve Ekker

The NIMH Director job announcement and search is back on: hr.nih.gov/jobs/search/... The NIMH Director provides executive direction and scientific leadership for NIMH, overseeing a $2.3 billion budget and ~1600 employees, trainees, and contractors. Up for the challenge 😉? Apply!

hr.nih.gov

22521

Steve Ekker @scekker.bsky.social · 20/08/2025

No, we are spending all of our leadership on even MORE gerrymandering.

Steve Ekker @scekker.bsky.social · 14/07/2025

Great to hear! I was lucky to have Dr Chris Wylie as a senior faculty member at UMN for years showcasing COB and Development. Really like the mission of COB

I used Grok to assess this, and Grok just 'spewed forth' - citations to primary sources. Terrible.

000

Indeed. Such BS when reviewers are making such demands. Let the authors decide on the story they want to tell!!!!

I’d think an ‘AI editorial tag’ could go a long way too

000

Maybe make it a tagged-annotated process. So even if issues are missed the first time, cannot correct later? Time to make our papers more like software where bugs just get fixed

110

Steve Ekker @scekker.bsky.social · 10/07/2025

When can AI tools dramatically reduce these costs while still keeping humans in the loop?

110

FishFolk - post if you are here in Madison!

030

Steve Ekker @scekker.bsky.social · 26/06/2025

@izfs.bsky.social Zebrafish Disease Models Society should be this handle too, but I do not think it's active. @zdms.bsky.social

110

Steve Ekker @scekker.bsky.social · 18/06/2025

A quiet conversation on the side, a closed-door phone call, a printout showing an email out of context... this 'whispernet' is a true poison in academia. Our 'culture of silence' feeds these voices, and the threat of being a recipient keeps most of the reasonable folk on the sidelines.

Carl T. Bergstrom @carlbergstrom.com · 07/06/2025

Reposted by Steve Ekker

Very smart, high ROI initiative from the Spencer Foundation. Providing bridge funding for canceled NSF grants: 1) directs funds to projects already screened by a top-noch review process. 2) avoids imposing proposal prep costs on researchers. 3) ameliorates the disruption from GOP sabotage.

spencer.org

Rapid Response Bridge Funding Program

In the face of recent abrupt shifts in federal funding for education research, including large-scale terminations of National Science Foundation (NSF) research grant awards, we have developed a rap...

6771333

Reposted by Steve Ekker

Mark A. Hanson @hansonmark.bsky.social · 03/06/2025

URGENT: FlyBase has lost practically all its funding overnight; even user fees are tied up in denied grant funding. 🤬🤯 Any lab using @flybase.bsky.social please donate using the link in post below. This incredible community, on whose backs our #Drosophila labs depend, can't be left out to dry.

4225253

Steve Ekker @scekker.bsky.social · 31/05/2025

Should have called it 'Just Keep Swimming'!

030

Steve Ekker @scekker.bsky.social · 15/05/2025

From a wonderful group of colleagues! We hope it helps with thoughts and suggestions on using #zebrafish. Ping me with any questions. #werallfish

Steve Ekker @scekker.bsky.social · 03/05/2025

Sebastian is awesome. Catch him while he has extra availability.

Steve Ekker @scekker.bsky.social · 03/05/2025

Genotyping zebrafish can be a challenge. We have been refining a non-destructive approach that is scalable, accessible and can be used in 48 well and 96 well formats. We use this for measuring mitochondrial DNA editing and heteroplasmy, but works on nuclear DNA, too. www.biorxiv.org/content/10.1...

biorxiv.org

Non-Destructive Larval Genotyping of Danio rerio for Mitochondrial DNA Genetics

The rapid advancement of nuclear and mitochondrial genomic editing tools has created an urgent need for efficient, non-lethal larval genotyping methods in zebrafish Danio rerio research. This study op...

040

Steve Ekker @scekker.bsky.social · 07/04/2025

Great thread. It’s getting even more complicated when genetic compensation is coming to light. Meaning there are additional layers of both genetics and environmental inputs that will change the system and our phenotypes. The ‘winners’ of simple views are those selling articles by clickbait headlines

000

Reposted by Steve Ekker

Brendan Evano @brendan-evano.bsky.social · 07/04/2025

🚨 JOB OFFER in our lab at Institut Pasteur 🚨 We are looking for a postdoc! Please repost 🙏

Postdoctoral research position in Institut Pasteur

22725

The most cited it the developmental staging series by Chuck Kimmel, as I recall.

120

The silk region of China is pretty amazing! Been growing silk worms for 4000 years…

130

Lastly, the #zebrafish field is in great shape! So many 'first-as-PI' talks from new faculty. Their energy, creativity and enthusiasm were contagious.

040

Finally - just as shown by the recent Olympic Games, Paris is an amazing host city for science! Thanks to all of the organizers for a great meeting,

While I appreciate the complex toolkits, some of the best talks are those just asking simple questions. How do fish sense bacteria in their environment? Super interesting that we still have new chemosensor cells to find.

The most impactful talks, to me, were those where old-met-new technologies and problems. scRNASeq combined with old-fashioned WISH and genetic lineages, plus germline genetics plus transient KD (CRISPANTs, morpholinos) - all in transgenic fish with tissue-specific reporters. Yes! 4D context!

scRNASeq is almost an obligate tool these days. While I am convinced it CAN be useful, it's being overused. But maybe AI tools are going to be able to fully mine this super interesting data. But speaking as the white-haired dude, I was struggling to find examples where scRNASeq yielded new genes.

I am personally convinced we will solve the mechanisms of regeneration in mammals from work in animals that have not lost this ability (ie zebrafish spinal cord neurons, heart). That will be in the next decade in detail. This will enable a new era of regenerative therapies.

040

$zebrafisharecool is still true! So, so much innovative work with this aquatic vertebrate. The tools and base understanding continue to enable fantastic new efforts from biochemistry to complex epigenetics.

What used to be 30 min (or 45 min) talks are now 15 min. Why? I think our attention span is shorter. But I think this is also a function of preprints. I did not have to hear ALL of the story if the full story was posted as a preprint. Preprints have just ended the time lag from talk to written word.

000

This is the first meeting I feel was at the same (or better) interactive as meetings pre-COVID. Just learned so much - was my privilege to learn from everyone!

000

First, preprints are here to stay. The top presented results were posted as preprints right before this meeting. My interpretation is that the meeting set the deadline for the writeups. This is a great synergy IMHO.

000

Steve Ekker @scekker.bsky.social · 04/04/2025

Just finishing up with the amazing 7th European Zebrafish Investigators' Meeting, hosted by the Paris Brain Institute. Some take-homes:

1191

Might have to do something special!