iiinova — where your digital self keeps evolving

Gabe Lynch @gpxl.bsky.social · 01/10/2026

The death penalty is cruel, sadistic, and immoral. The state should never have the authority to execute someone and it is enraging that far too many treat this as normal. Good God.

000

Reposted by Gabe Lynch

The Great Pumpkin Papers 🎃🎃🎃 @professormusgrave.bsky.social · 30/09/2026

It certainly seems like material explanations can’t explain all of voters’ behaviors—but what if there were another motivation?

3875

Reposted by Gabe Lynch

John Hatch @johnhatch.bsky.social · 30/09/2026

White Sox have been blessed by the Pope. Astros quaking in their cleats.

A letter from the Pope to the Chicago White Sox.

271

Reposted by Gabe Lynch

Joe Mason @moreorloess.bsky.social · 21/09/2026

A couple of my dad's black and white photos of abandoned buildings and towns on the Great Plains. I don't have locations, except the one on the right looks like the Nebraska Sandhills.

Black and white photo of an abandoned main street in a small town, mostly brick commercial buildings, weeds coming up from the sidewalk.

An old car, a shed, and an abandoned house in a grassland setting, maybe the Nebraska Sandhills.

4283

Gabe Lynch @gpxl.bsky.social · 18/09/2026

Does anyone have recommended reading on why AI doom is NOT the likely outcome, ideally from AI researchers? To be clear this is my belief (exceedingly unlikely) but given the flood of AI = skynet articles this past week, it would be good to have some references when this invariably comes up.

110

Gabe Lynch @gpxl.bsky.social · 18/09/2026

Nothing Compares In Comparison 2 U

110

Gabe Lynch @gpxl.bsky.social · 18/09/2026

I am curious: how do the following relations semantically differ to you when reading an academic paper (if they differ at all): 1. ~ 2. ≈ 3. ≃ 4. ≅

000

Gabe Lynch @gpxl.bsky.social · 16/09/2026

just logged on, is there a specific reason everyone is talking about rationalist cults / aella / the insanity of the agentic class?

000

Gabe Lynch @gpxl.bsky.social · 11/09/2026

I have never titled a paper in reference to the Beatles but did once give a talk with title

The text "Negative M_nu Mystery Tour" with the iconic Beatles rainbow lettering

131

Reposted by Gabe Lynch

Terence Tao @teorth.bsky.social · 11/09/2026

A group of 25 Fields Medalists, including myself, have made a joint declaration on Math and AI: mathandai.org . We welcome additional signatories. See also this article in the Economist announcing the declaration: www.economist.com/science-and-...

mathandai.org

Declaration — Math and AI

Read the declaration and add your name.

422052929

Gabe Lynch @gpxl.bsky.social · 11/09/2026

This is an interesting study but man it sucks always wondering, in the back of my mind, if people will ever mistake any of my writing for LLM produced stuff because it has the 'wrong words'.

010

Reposted by Gabe Lynch

Ancient Egyptian Prayer Bot @egyptianprayerbot.bsky.social · 10/09/2026

I am Yesterday, I know Tomorrow. (BDB)

03715

Gabe Lynch @gpxl.bsky.social · 10/09/2026

if AGI superintelligence poses a threat to humanity, and most future intelligences are therefore artificial, then what are the odds that we are interacting with the very first few AGIs? The Bayesian priors are not very favorable.

000

Reposted by Gabe Lynch

CosmicRami 🏳️‍🌈🏳️‍⚧️ @rami.spaceaustralia.com · 09/09/2026

The "$22M in computing power" is a metric. That kind of money, that these orgs have, could have been spent on 440 PhDs @ $50K each. If this truly was about solving a problem for maths, for humanity, for whatever, 440 PhDs would have had far greater scale of global impact.

1387

Gabe Lynch @gpxl.bsky.social · 09/09/2026

hit dog: and another thing: im not hollering. please dont put in the newspaper that i hollered.

000

Reposted by Gabe Lynch

Scientific American @sciam.bsky.social · 08/09/2026

Prior to OpenAI's announcement today that its AI solved one of the seven Millennium Problems, rumors were circulating online among mathematicians over the proof's origins—OpenAI has denied all allegations.

scientificamerican.com

OpenAI claims blockbuster math breakthrough amid swirl of controversy

Prior to the announcement, rumors were circulating online among mathematicians over the proof's origins—OpenAI has denied all allegations

42012

Gabe Lynch @gpxl.bsky.social · 08/09/2026

Posting this without any discussion of Tristan Buckmaster’s statement on what occurred behind the scenes here is journalistic malpractice. writtenhttps://cims.nyu.edu/~tristanb/statement.pdf

cims.nyu.edu

12510

Reposted by Gabe Lynch

Max Kennerly @maxkennerly.bsky.social · 08/09/2026

... so there are Navier-Stokes draft solutions out there, and also it seems like OpenAI is spying on mathematics work in GPT, copying approaches, and then throwing their massive compute resources at others' ideas to publish it first. Statement from Andrew Curran: cims.nyu.edu/~tristanb/st...

Today, Levent Alp¨oge and I have made public three results: finite-time blowup with smooth forcing for incompressible porous media, for Boussinesq, and for 3d incompressible Euler. We believe we also have blowup for hypo-dissipative Navier-Stokes. We are not releasing that paper today: unlike the above, the Lean verification has not yet finished. We do not yet have anything resembling a presentable writeup. I mention it because it is suggestive of a path to unforced Euler. The program this fits into was not started by us nor was it proposed by a Large Language Model. The credit for the basic idea of this program goes to Diego C´ordoba and Luis Mart´ınez-Zoroa, who for several years have been exploring the construction of forced blow ups. We took their work as a starting point, using Large Language Models to push their program to completion. Concretely, what Levent and I did was to take the C´ordoba and Mart´ınezZoroa program, which achieved blowup results with rough forcing, and, with a great deal of help from LLMs, push it to smooth forcing and to the incompressible Euler equations. The ideas making this line of attack possible are due to C´ordoba and Mart´ınez-Zoroa. Let me make plain what I have said to colleagues in private: in view of this body of work, I believe Luis Mart´ınez-Zoroa deserves a Fields Medal. My work with Levent has been a purely personal collaboration, free of any institutional agreements or official involvement by either of our employers. We used several LLMs throughout: Anthropic’s Claude, OpenAI’s Codex, especially with GPT-5.6 Sol and, more recently, Astra. The latter was only used for writeups and auditing our arguments.

He replied the same day: “If you are willing to give any details it would be useful to avoid competing here and in general we are always thrilled when mathematician make progress with our models. Additionally if there is anything in terms of compute from OpenAI’s end we would be happy to provide it.” I asked to speak the following week. On Friday, September 4th, I was asked whether I could meet that day; I again said the following week. At 12:45 on Sunday, September 6th, I was asked whether I could meet “at any point today.” Sebastien Bubeck joined. The three of us spoke twice that afternoon. Levent was not on the calls. I was told that an internal OpenAI model had produced a proof of finite time blowup for the forced Navier-Stokes equations. When Levent asked by text for

the precise statement, the answer was: “Existence of forced blowup in R3 and T3,” with “the forcing function is smooth option c and d in Fefferman.” I was told the proof is about 100 pages. I have not seen it. I should say here why I interpreted their statement the way I did, the interpretation I will discuss below. The route to the Clay problem through a smooth force, options c and d in Fefferman’s statement of the problem, is the route Luis and Diego opened and the one Levent and I had quietly chosen to attack. Almost nobody else I know of was working on it. It is not the direction one arrives at in a few days by giving a model the problem statement. When I heard “forced,” it was a bright red flag. I was shown a prompt and told the internal research model had simply been given the problem statement. Levent had been told by Sebastien “very little human input” had been used. This turned out not to be true. Over the course of the call, as members of their team sent Sebastien corrections and details over their internal chat, it emerged that an entire team had been working on the problem, that this was one of a number of things that was tried, that work had started on the unforced problem, that the team first set the model on easier problems, including Euler, that even the prompt that had been shown to me had been written by prompting Codex, and that an insane amount of compute had been used. I asked when the first prompt had been sent by them. This question was not answered directly by OpenAI for some time. Eventually it was agreed that it had been sent in the past few days, after information about our work had reached OpenAI. I asked whether the model had been trained on, or had access to, our sessions in Codex, into which we had been putting all our drafts for the whole of this project. I was told the model did not look up user data. I asked again, about training, and I did not get an answer.

27616

Gabe Lynch @gpxl.bsky.social · 02/09/2026

getting a "I believe we were visited by WIMP dark matter on June 16 2023" bumper sticker

020

Reposted by Gabe Lynch

Chris Mihos @chrismihos.bsky.social · 30/08/2026

Ahh, I'm seeing a lot of particle physicists riding into astronomy research, ready fix our field and solve all our science questions for us. I mean, how hard can astronomy be, right? Welcome in, dear.

1950s man comes home, where his wife and children are hiding behind the door holding carving knives. His wife is labelled "Systematic uncertainties" and his kids are labelled "Non-Gaussian Errors" and "Dust".

1512817

Gabe Lynch @gpxl.bsky.social · 22/08/2026

it is amazing how this administration is genuinely attempting to make every good thing worse

011

Gabe Lynch @gpxl.bsky.social · 20/08/2026

imagining when my cursor goes from one monitor to another its like going through a spacetime jumpy and it emerges disoriented and confused on the other screen. godspeed little arrow, godspeed.

010

Reposted by Gabe Lynch

Boaz Nash @boazhsan.bsky.social · 19/08/2026

Even Einstein died with a messy desk

photo of Einstein's desk on the day he died

032

Reposted by Gabe Lynch

Gabe Lynch @gpxl.bsky.social · 18/08/2026

Wikipedia is so charming

a screenshot of the wikipedia page for Diagoras of Melos showing the pencil drawing uploaded by a Wikipedian to illustrate an anecdote.

a screenshot of the file metadata for the pencil drawing image showing it is the user’s own creation

052

Reposted by Gabe Lynch

Alexandre Adler @bachpropagate.bsky.social · 13/08/2026

Missed this when it came out, worth a read if you're playing tau.

011

Reposted by Gabe Lynch

Aaron Rupar @atrupar.com · 12/08/2026

Peggy Flanagan: "We need to rip ICE apart and stop from them from terrorizing our communities"

339196393905

Reposted by Gabe Lynch

Lake Michigan! Live! @lake-michigan.live · 12/08/2026

Current* conditions near Milwaukee, WI:

Southern view from the WITI broadcast tower in Estabrook Park overlooking downtown Milwaukee and the shoreline of Lake Michigan. // Image captured at: 2026-08-12 04:36:04 UTC (about 1 min. prior to this post) // Image sourced from: fox6now.com // Current Temp in Milwaukee: 71 F | 21 C // Precip: clear sky // Wind: S at 4 mph | 6 kph // Humidity: 91%

0101

Reposted by Gabe Lynch

Big Wonderful Wyoming @wydot.bsky.social · 07/08/2026

Cody, 8/7/2026, 4:58:08 PM

0252

Gabe Lynch @gpxl.bsky.social · 05/08/2026

me trying to reduce the dimensionality of my phenomenological model

020

Reposted by Gabe Lynch

Gabe Lynch @gpxl.bsky.social · 04/08/2026

New paper out today, so I'm making a really short #cosmology thread. This work was led by Isaac Sierra here at UC Davis but I was happy to contribute! We looked at whether exploding primordial black holes could explain a mild disagreement between CMB and BAO data... 1/n arxiv.org/abs/2608.01919

arxiv.org

Boosting the optical depth to Thomson scattering with primordial black hole evaporation at high redshift

BAO and CMB data are somewhat discrepant when interpreted in the context of $Λ$cdm, discrepancies that show up as a `matter density deficit' and as a `CMB lensing excess'. One possible resolution is a...

161

Gabe Lynch @gpxl.bsky.social · 04/08/2026

arxiv.org

Boosting the optical depth to Thomson scattering with primordial black hole evaporation at high redshift

161

Gabe Lynch @gpxl.bsky.social · 29/07/2026

taking a lot of self control to not comment on some FB posts discussing the Hopkins bike lanes

210

Reposted by Gabe Lynch

Jean Walsh @therealjeanwalsh.bsky.social · 29/07/2026

SAFE HOPKINS!!! This is huge. Thanks to Berkeley Mayor and Council for their visionary leadership and to everyone who has worked so hard on this for so many years.

2335

Gabe Lynch @gpxl.bsky.social · 29/07/2026

berkeley city council bumping tonight!

a photo of a packed public council meeting

110

Reposted by Gabe Lynch

lauren @lauren.rotatingsandwiches.com · 28/07/2026

say what you will about AI, always-on voice recording and auto dictation has been a game changer for my workflows. Save post and close app. Yeah Footlong meatball sub on white. No veggies. Don't toast it. Ranch and barbecue sauce. Yeah, ranch and bargecue sauce. More. More. Little more please

6186181470

Reposted by Gabe Lynch

Alexandre Adler @bachpropagate.bsky.social · 28/07/2026

Sorry, last one. 🚀 🌌

A four panel (square) reaction image where Peter Parker (as played by Tobey Maguire) puts on his glasses in two panels on the left, and picture on the right becomes clear. Top right is the SPCX ticker going down. Bottom right is the CMB TT spectrum plotted in log scale: the two curves look similar with a bump at the beginning and a series of wiggles trending downwards afterwards. If you have questions about the physics of the curve or the CMB please don't hesitate to ask!

142

Reposted by Gabe Lynch

Walk Bike Berkeley @walkbikeberk.bsky.social · 27/07/2026

Council votes on Hopkins tomorrow 7/28! Tell them to make the street smooth and safe for all! 👇

0179

Reposted by Gabe Lynch

Berkeley Neighbors for Housing and Climate Action @bnhca.bsky.social · 28/07/2026

In case you need a reminder, tomorrow’s Council meeting is CRITICAL and safe biking will need a lot of support - ideally in person!

183

Reposted by Gabe Lynch

Alexandre Adler @bachpropagate.bsky.social · 27/07/2026

Hey wait a minute. 🔭

The Planck CMB temperature power spectrum.

4285

Reposted by Gabe Lynch

Idothethinking @idothethinking.bsky.social · 27/07/2026

The battle of Hopkins bike lanes are back, the final battle, and I got a lot to say. I haven't attended a Berkeley city council meeting in 7 years, but on Tuesday I'll be there, and I hope you will too. darrellowens.substack.com/p/the-final-...