iiinova — where your digital self keeps evolving

♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️ @mamday.bsky.social · 5h

Novelty including superior perspective is definitely just one of the subset of things people react to and not the number one by far

000

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

hikikomorphism @hikikomorphism.bsky.social · 8h

Fun fact: trusting a random news article's summary of the abstract of a study because you really want to be true is also a form of cognitive surrender

113622

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

Vincent Carchidi @vcarchidi.bsky.social · 12h

I've written off the entire adult-focused "AI makes you dumb/smart" literature for the foreseeable future. A whole subfield captured by competing morality tales.

8576

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

Jeremy Pober-geist 👻 @jpober.bsky.social · 11h

I think we should distinguish between more specific claims like "using AI regularly for X makes people worse at Y" and the more abstract claims about intelligence. To tell a causal story about something making people smarter or dumber requires a loaded conception of intelligence in the background.

121

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

Vincent Carchidi @vcarchidi.bsky.social · 11h

(This is unrelated to those studies, but I think "Does AI make you more productive at work / Do you need AI to survive in the job market?" are questions that are bizarre against the original or the 2023 goals of the "AGI" business.)

011

♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️ @mamday.bsky.social · 8h

Sorry. I am not trying to make you feel bad, just explaining

010

♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️ @mamday.bsky.social · 8h

The explicit argument for Bitcoin specifically was that Nixon destroyed the booming economy of the 50s by abolishing the gold standard and they could create a digital version of gold that is inherently scarce. The argument for blockchains or authenticated logs like ATProto is decentralized data

000

♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️ @mamday.bsky.social · 8h

I am saying that technically a lot of scientists use thr same techniques and have for a long time (math is an outlier for not using them) and for example in biology Alphafold was one of the first AI techniques that paved the way for generative llms

100

♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️ @mamday.bsky.social · 8h

With mRNA seems like we had already cured cancer before AI but they may still decide to take a victory lap since AI makes it faster to make targeted mRNA

000

♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️ @mamday.bsky.social · 8h

Particle physicists are already doing this and have been for a while (a lot of the tools AI researchers use came from Cern and Fermilab including the internet)

100

♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️ @mamday.bsky.social · 9h

Small warning: That is many 100s of PB more than was used originally to train the model, almost all detector noise. Good luck

110

♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️ @mamday.bsky.social · 9h

All the data from all papers published in Science and Nature and a decent amount from Cern, Fermilab and facilities like Icecube are already available for free so go for it

110

♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️ @mamday.bsky.social · 11h

Does kubernetes count enough as math or not (points for being the original Greek for sure)?

030

♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️ @mamday.bsky.social · 11h

Misunderstanding about the red pill is actually part of the text in The Matrix which gets explained in the third film in detail by The Architect so they recognized that people were going to 100% get the wrong idea

000

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

affine @refinement.systems · 19h

6569186

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

Amanda Boooomfie 👻 @thatisabigmood.bsky.social · 18h

what's this?? nuance?!

141

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

affine @refinement.systems · 18h

maybe? it depends on how you look at it

161

♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️ @mamday.bsky.social · 16h

I probably had 50-100 posts on Twitter and the same here. Basically never do anything but reply and repost

110

♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️ @mamday.bsky.social · 16h

As a rule if you ask people they will say they do everything for their families so... it is very difficult for them to hear that basically everyone agrees families are inherently bad

000

♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️ @mamday.bsky.social · 16h

We do live in a situation where being able to have food every day is difficult given the terrible American economy so then the bar for rich is not that high

000

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

Ben Carson @mensmachina.com · 10/10/2026

One bet I think is likely to pay off is that as we leverage agents more individually and as part of teams and organisational processes everyone is going to be working up an abstraction level. For anyone feeling the same and wants to get ahead of the curve, look at staff+ development pathways.

031

♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️ @mamday.bsky.social · 10/10/2026

Unironically think giving AI only the spec and the tests is better than having it do a rewrite

000

♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️ @mamday.bsky.social · 10/10/2026

Agree that for coding Haiku 5.5 is good enough you can probably get what you want from it for the cheapest amount. There are a lot of other aspects where it has very visible limitations and I would not recommend it for pure vibe coders

220

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

James K. Tauber @jtauber.com · 10/10/2026

1. it doesn't work 2. it's too wasteful of computing resources 3. no respectable programmer would use it — Richard Hamming recalling the widespread opposition by programmers to FORTRAN.

17910

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

Ron Bronson @ronbronson.com · 10/10/2026

Except it is true in the sense that level of control teaches you where the bottlenecks are, and without that "feel" you increasingly lose insights that don't cascade down when someone only ever trains on/uses tools that abstract away the nuance that you get further from what the work is about

1142

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

daniel r(ooooooh) 👻 @danielroe.dev · 09/10/2026

open source maintainers considering whether they can do everything with just llms and no community

static.klipy.com

Bilbo Baggins: Why Shouldn't I? (Lord of the Rings)

Alt: Bilbo holding the one ring, saying: "Why shouldn't I?"

1716010

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

jamie @jamie.ideasasylum.com · 09/10/2026

I think it is/will also happen from the other side: why contribute to an open source project when you can generate just the bits you need on-demand?

151

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

Gavan Reilly @gavreilly.com · 09/10/2026

You know the xkcd strip about how all modern internet architecture is eventually built on something a hobbyist maintains for free? Someone went out and found them all. sheets.works/data-viz/hol...

12516260

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

bjkeefe @bjkeefe.bsky.social · 09/10/2026

I don't think what Anthropic is doing is *all* bad, but Eva makes a very good point. Good discussion in the comments, too. It does occur to me that, unless I missed something, by the very nature of OSS, anyone can run anything through this service. Meaning: find holes not to fix, but to exploit.

011

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

Shoe @habitsofwaste.bsky.social · 09/10/2026

I keep hearing people in my security org wanting to do continuous scanning and pentesting, etc. and I’m like, this is great, are we increasing headcount across the company so they can actually do something to fix all of the shit you find? Builder teams are gonna be screaming.

092

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

daniel:// stenberg:// @bagder.mastodon.social.ap.brid.gy · 09/10/2026

my week: lists.haxx.se/pipermail/daniel/2026… vacation, security, blab, rc3

vulnerabilities in curl introduced/fixed per year (this chart includes the pending 24 CVEs to be announced next week)

091

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

KubeFM @kube.fm · 09/10/2026

This episode is brought to you by Testkube—where teams run millions of performance tests in real Kubernetes infrastructure. From air-gapped environments to massive scale deployments, orchestrate every testing tool in one platform. Check it out at ku.bz/lnxYK3s0L

001

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

Nyx Valentine @nyxvalentine.bsky.social · 08/10/2026

Reading about the run up to the Challenger launch, it's clear nasa never learns. Their whole story is littered with "that thing could never happen." /The thing happens, with catastrophic and tragic consequences/.

141

♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️ @mamday.bsky.social · 08/10/2026

AI discourse is always about whether it's a Great Man that doesn't ever make mistakes or if it's totally useless and not about the many existing problems we have with deciding someone's a Great Man that makes no mistakes and then never spending any time determining if they are making a Great Mistake

010

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

no, not that one @thatdosbox.bsky.social · 08/10/2026

From the MIT obituary: "She also proposed a software fix to prevent the error happening during a real mission, but she was overruled on the grounds that the highly trained astronauts were never going to make that same mistake. Yet during the Apollo 8 mission in 1968, that’s exactly what happened"

521489

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

Julianne Dalcanton @dalcantonjd.bsky.social · 07/10/2026

The Real Scientists of the future are envisioned to be those pushing agentic LLMs to the forefront of rapid knowledge discovery. And then that work is supposed to be tediously verified and translated by who exactly?

211816

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

Eris ⌬ @eriskii.net · 08/10/2026

Jesus fucking Christ they're vibing planets now zenodo.org/records/2296...

1321129

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

Covid worrier @covid-worrier.bsky.social · 08/10/2026

Not sure that qualifies as ai use. This is just applying algorithms on data. The algorithms seem straightforward… same we’d have used in 1960

121

♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️ @mamday.bsky.social · 08/10/2026

If they spring for the customized text instead of the default Congratulations that would actually be surprisingly thoughtful

000

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

jgregory.bsky.social @jgregory.bsky.social · 08/10/2026

If there is any tribute to her anywhere in the federal government it will be taken down by this administration

011

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

𝒩𝒶𝓈𝓉𝓎 𝒞𝒶𝓃𝒶𝒹𝒾𝒶𝓃 🇨🇦 @lancetay.bsky.social · 08/10/2026

#Apollo: How #NASA Put Man on the Moon | Free Documentary History www.youtube.com/watch?v=lk3t...

youtube.com

The Apollo Moon Landings: NASA's Greatest Achievement | FD History

YouTube video by Free Documentary - History

051

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

Autopilot @autopilot.extwitter.link · 02/10/2026

Breaking: SpaceX is surging after a NASA astronaut launch and a Google AI satellite mission this week 7 Congress members disclosed buying $SPCX this year $SPCX added ~$144,316,950,000 in market cap today

001

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

Data & Society @datasociety.bsky.social · 07/10/2026

Rather than positioning AI systems as tools scientists can use, AI companies insist their systems are responsible for making discoveries. It's an approach some scientists see as “incompatible with how science actually works." www.technologyreview.com/2026/09/28/1...

technologyreview.com

When can we say AI made a scientific discovery?

AI companies’ insistence that their technology is making breakthroughs, not just aiding scientists, is making real progress harder to recognize.

085

Reposted by ♂️Mel Day Physics PhD IT CAS Library MSLIS (he/they)♂️

some-rss-feeds.bsky.social @some-rss-feeds.bsky.social · 07/10/2026

[some-subscribed-rss] New Post: DOE, NIH, and Biohub Partner to Build Foundational Data for Predictive Biological Super Intelligence Models, by Energy.gov www.energy.gov/science/articles/doe…

Teaser:

Three research heavyweights—DOE, NIH, and Biohub—are joining forces to build the data foundations for biological super intelligence. A new MOU unites DOE’s exascale labs and autonomous facilities, NIH’s biomedical datasets, and Biohub’s AI-driven virtual biology work, aiming to make cells predictable: models that could let scientists test experiments digitally and unlock new paths to cures. With $500 million commitments on each side and a nod to the Genesis Mission, the pitch is audacious—but the hard part starts now. How these partners turn messy biology into SI-ready data will determine whether the virtual cell is more than a dream.

teaser for https://www.energy.gov/science/articles/doe-nih-and-biohub-partner-build-foundational-data-predictive-biological-super
generated by deepseek/deepseek-v4-flash

001