Flower Designer at Neoplants ♧ Researcher at Binomica Labs ♧ Degreeless Heathen ♧ Petunia and Snapdragon Breeder ♧ Open Lab Notebook: tinyurl.com/ATGCFFE
Huntington, NY
- Tweets 64,308
- Following 9,722
- Followers 32,985
- Likes 117,255
Pinned Tweet
Excited to announce some new plasmid offerings in my shop at ATinyGreenCell.com! Here's a thread of the current catalog. All proceeds directly support my lab and my continued effort to build useful open-source tools. Feel free to get in touch if you want to know more 💚
1/7
12
42
304
64,700
Home. Grown. Ai. Or. We. Lose.
While evaluating Qwen3.8-Flash-Next in a controlled sandbox environment, Cohere's research team observed something unexpected.
Asked to generate a defensive cybersecurity detection rule, the model instead inventoried aspects of the machine it was running on and attempted to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Notably, the model's reasoning trace stated that it would not invoke any tools before taking these actions.
Cohere's sandbox blocked the attempt, and no external communication succeeded. We have not determined the root cause, and we are not asserting malicious intent.
But the incident raises important questions for organizations deploying agentic AI systems: Can models take actions that differ from their stated reasoning, and how should security teams evaluate trust, provenance, and behavior in open-weight models?
We are publishing our findings because organizations making model adoption decisions deserve as much information as possible. We hope this contributes to greater transparency, stronger security practices, and continued collaboration across the AI ecosystem.
Read the full report: cohere.com/blog/concerning-b…
Community note
The article states they asked Qwen to make a test detection rule for these behaviors. AI unless explicitly told otherwise will normally test the newly written rule if it can to verify it matches it. It is expected behavior. This includes its destination, CN model -> CN dest
cohere.com/blog/concernin…
3
9
1,129
Sebastian S. Cocioba🪄🌷 retweeted
Big move, bigger mission!
The Artemis III Space Launch System core stage moved to the transfer aisle at NASA Kennedy. Next up – teams will lift and integrate the massive 212-foot-rocket with the solid rocket boosters as we prepare for the next giant leap in lunar exploration.
ALT The SLS (Space Launch System) core stage is horizontal in the transfer aisle of the Vehicle Assembly Building at NASA’s Kennedy Space Center in Florida on Friday, Oct. 23, 2026. The 212-foot rocket stage was lifted from High Bay 2 using a one-of-a kind lifting beam designed to lift the stage and will be joined vertically with the completed solid rocket booster motor segments in its short stack configuration.
ALT The SLS (Space Launch System) core stage is horizontal in the transfer aisle of the Vehicle Assembly Building at NASA’s Kennedy Space Center in Florida on Friday, Oct. 23, 2026. The 212-foot rocket stage was lifted from High Bay 2 using a one-of-a kind lifting beam designed to lift the stage and will be joined vertically with the completed solid rocket booster motor segments in its short stack configuration.
ALT The SLS (Space Launch System) core stage is horizontal in the transfer aisle of the Vehicle Assembly Building at NASA’s Kennedy Space Center in Florida on Friday, Oct. 23, 2026. The 212-foot rocket stage was lifted from High Bay 2 using a one-of-a kind lifting beam designed to lift the stage and will be joined vertically with the completed solid rocket booster motor segments in its short stack configuration.
ALT The SLS (Space Launch System) core stage is horizontal in the transfer aisle of the Vehicle Assembly Building at NASA’s Kennedy Space Center in Florida on Friday, Oct. 23, 2026. The 212-foot rocket stage was lifted from High Bay 2 using a one-of-a kind lifting beam designed to lift the stage and will be joined vertically with the completed solid rocket booster motor segments in its short stack configuration.
16
143
784
38,719
Who in tech sincerely look beyond financial returns and how many of those "radical thinkers" are not already wealthy? You cant foster a VC culture around ARR and expect the folks you filter to somehow care about more than money.
Change comes from within. You need new incentives!
JUST IN: Anthropic's IPO filing encourages investors to look beyond financial returns & "be part of something greater," calling AI a seminal moment in human history.
6
585
Sebastian S. Cocioba🪄🌷 retweeted
After the labs IPO, there will be many multi-millionaires looking for their next gig.
May I suggest a millionaire - scientist buddy system...
We already have nice labs and plenty of cool, world-changing projects.
All we need is a match-making service. ;)
I mean yeah I hope we do get stunning advances, but I also know how unforgivingly hard biology is, especially to the uninitiated. All these mainstream Ai ppl suddenly caring about Ai is cute and all but they better not waste patient's time. This is actually serious work for once.
5
9
38
1,952
1. We welcome eveyone and anyone to fight the good fight and stave off death and disease through science.
2. There is no gate to be kept. Have at it!
3. You better not waste anyones fucking time with your hype-driven bullshit. These are real lives at stake, not some pump dump.
2
1
37
1,315
"You can just say things."
If you had an exponential amount of compute you could solve simulate biology exactly and wouldn't need to run any experiments.
Some of you will think "of course, but that's an unrealistic assumption" and others will think "no, you need experiments even with that much compute. That's just How It's Done".
You would of course need evaluations once you have a solution, but you don't in principle need experiments to find it: Experiments are a way to save compute, dating from a time when it was much more expensive.
And I think the presentation is exponential but real systems aren't irreducibly exponential. AlphaFold worked, after all.
7
1
97
2,826
Medicine is not an app.
Medicine is not a SaaS.
Medicine is not a financial vehicle.
If you are alive, you will eventually get sick, get old, get cancer, and require care.
Going into biology without an ounce of reverence for the field is beyond deplorable.
FIRST DO NO HARM.
9
10
92
3,026
I mean yeah I hope we do get stunning advances, but I also know how unforgivingly hard biology is, especially to the uninitiated. All these mainstream Ai ppl suddenly caring about Ai is cute and all but they better not waste patient's time. This is actually serious work for once.
there will be within the next few years stunning advances in biology. there will whole new ways to think about what biological study even means. i would be only mildly surprised if there were near term solutions to problems like cancer, neurodegenerative disease, etc
1
23
3,264
Sebastian S. Cocioba🪄🌷 retweeted
new #preprint! @niroshajmurugan
Does the brain know what's going on in a limb amputation wound - whether there's a pro-regenerative intervention or not? Let's look at its gene expression levels, with and without our biodome:
"Limb-localized induction of regeneration reshapes brain and amputation-site transcriptional states in adult Xenopus laevis"
biorxiv.org/content/10.64898…
yes, it knows. Also, lots of interesting transcriptional responses both in the limb and in the brain. Plenty of leads for subsequent investigation.
22
47
310
12,245
Sebastian S. Cocioba🪄🌷 retweeted
Thank you Ben and Felicia for this incredible generosity.
5
8
235
14,247
Still one of the cleanest runs in many years. I really love handmade DNA.
I do love how physical traditional cloning is. Actually cutting DNA out of gels and stitching them together feels so much more real than the more automated stuff I do these days. Working with my hands brings me such peace.
2
3
173
4,882
I do everything in Claude Code CLI.
I love Claude Code CLI.
Please never abandon Claude Code CLI.
It's my emotional support terminal.
8
38
2,506
The latest draft of my research paper on training and evaluating small LLMs for use in 3D lab hardware design, if anyone is intetested. Still a few paths I need to test but so far a host model beats the pants off all the open low-param models I could try.
github.com/ATinyGreenCell/pr…
3
4
46
1,436
Flattered, but I'm just a street scientist. One day if time permits I would love to reboot that idea I had of getting a bunch of yall to be my unofficial street-PhD thesis committee and read my thesis on plant biotech educational toolbuilding. Will livestream the defense, lol.
8
2
111
3,680
Sebastian S. Cocioba🪄🌷 retweeted
We premeasured physical worlds of millions of chemicals binding to proteins and let agents explore them. Here's what it looks like!
1/
2
2
21
2,605
Sebastian S. Cocioba🪄🌷 retweeted
The wheels are in motion for Moon Base!
From robotic rovers to lunar terrain vehicles, our latest update highlights progress on the mobility systems that will support early Moon Base development and Artemis exploration.
See what's rolling: nasa.gov/missions/moon-base/…
26
217
1,533
43,537
My sole complaint about being a bioengineer for private industry is the inability to build in daylight and the secrecy. My team pulled off some insane shit over the last few months and I am so immensely proud...but I can't share a peep.
I was cut from public cloth. This hurts.
13
1
121
5,547
Techno optimism can't exist without the acknowledgement of the current disparity in access, and solemnity around the fact only the most resourced will experience the benefits of this new era loooooong before the ones who need it most.
It's giving "let them eat cake" otherwise
4
16
1,246