Jaffray Woodriff

Chief Scientist & CEO QIM. UVa School of Data Science (Founder/Funder). Embrace Simplicity, Beware Complexity. Coding at CODE. Honesty. Sleep. Squash.

[Andrew Gordon Wilson @andrewgwils]

I remember in biology being told "viruses are not alive" as if it were a profound revelation. But we excluded them based on counterintuitive criteria such as independent replication. It's because "we said so". Many seeming profundities can be traced to arbitrary definitions.

[Jordan Guy🏳️‍🌈🏳️‍🌈 @GuyJordan]

The irony here is that America was literally founded on secularism. The Founders had just watched centuries of Europe tear itself apart with religious wars — Catholics vs. Protestants, Anglicans vs. Puritans, you name it. Their solution? A Constitution that kept government out of religion, so people could worship (or not) freely. That’s why we don’t have a Church of America like England had a Church of England.

Christianity has absolutely shaped our culture, but the genius of the American system wasn’t to enshrine one faith — it was to protect all faiths by refusing to elevate any. Madison called it a ‘great barrier’ against tyranny. Jefferson said the First Amendment built a ‘wall of separation’ between church and state. Washington himself reminded us the government isn’t competent to run souls — just civil order.

So if things feel like they’re going ‘downhill,’ maybe the problem isn’t secularism. After all, some of our greatest moral victories — abolition, civil rights, even women’s suffrage — were achieved not because government enforced a religion, but because our secular Constitution made room for people of faith and reason to push for justice.

In other words: secularism wasn’t America’s downfall, it was the foundation. Blaming it is like blaming the seatbelt for the car crash.

[Clifford Asness @CliffordAsness]

What’s missing is the part of your cerebral cortex that understands that the #s needed are orders of magnitude off. Tariffs can’t replace income tax unless you shrink government by an utterly massive amount (and without touching entitlements, the right’s new position, it’s beyond impossible — actually it’s already beyond impossible this just makes it a joke). Your motto is “think for yourself.” Try it. Don’t just repeat nonsense and lies.

[Jeff @JeffreyPeterMN]

He’s said it so many times now, it’s gotta be deliberate. Big numbers simply sound more impressive to people who lack critical thinking ability. He’s a good salesman. Let’s be honest, it’s impressive this guy got his “conservative” base to get on board with a consumption tax (tariffs) that hits many of them hardest.

[Andrew Gordon Wilson @andrewgwils]

The “bitter lesson” asserts we have a cognitive bias towards encoding human knowledge over computation. Ironically, we also have a bias towards uncritically accepting what seems like bad news as truth — a “bitter lesson”, or “no free lunch” theorems.

[John Palmer @johnpalmer]

@sama I believe we can do better than all three.

[John Arnold @johnarnold]

Someone should write about how California's ban on employee noncompete agreements has spurred faster progress, greater innovation, and stronger competition in AI, benefitting employees and consumers alike while companies thrive, and what it means for the broader national debate.

[gfodor.id @gfodor]

The optimal amount of hallucinations is not zero. How do people not understand this. The complainers are midwits. Einstein, Tesla, Ramanujan all almost certainly experienced the kind of “hallucinations” we see in models. Snuff those out entirely and you will never reach ASI.

[Harlan Stewart @HumanHarlan]

Quick impression of GPT-5 announcement: seems more about making a more useful product than about a leap in raw capabilities

[John Arnold @johnarnold]

Blue-state Republicans and red-state Democrats are the most moderating influence in Congress and are usually instrumental to any bipartisan legislation. Their continued erosion through gerrymandering will only make Congress more polarized and less capable of governing.

[Gurwinder @G_S_Bhogal]

@paulg Even better, find a kind of work that you're so interested in that you'll learn AI to do it better than everyone else can.

[Andrew Gordon Wilson @andrewgwils]

Something I noticed at ICML: for the most part, VCs are strikingly disinterested in the technical content of the conference, and the discipline at large. You could be describing what obsoletes the transformer, and they couldn’t care less.

[Dwarkesh Patel @dwarkesh_sp]

Every time someone proposes training a model for some domain (virtual cells/materials/whatever), the first sentence should be: input data: x, output data: y, loss function: z.

Would make it a lot easier for dum dums like me to understand what exactly is being proposed.

[Jaffray Woodriff @JaffrayW]

https://t.co/fudJsrrHgK

[Jaffray Woodriff @JaffrayW]

Heya @grok based on the kind of content I enjoy who should I follow that I'm not currently?

[Andrew Gordon Wilson @andrewgwils]

Perhaps a majority of brilliant ideas, near their conception, are not only dismissed, but often ridiculed. My understanding is that early versions of GPT were written off out of hand by supposed luminaries (who now wholeheartedly embrace and replicate the technology). 1/2

[Jim Rutt @jim_rutt]

"One of the most basic observations we can make about the world is that the vast majority of people act on scripts—essentially roles that have been defined in advance."

- @SamoBurja on his distinction between live and dead players

[Mason @webdevMason]

"Homeschooling" is so much more heterogenous than schooling that it's trivially easy to find homeschooled adults who seem either damaged or superpowered as a direct result of it

[Andrew Gordon Wilson @andrewgwils]

The no free lunch theorem is irrelevant.

[John Maeda @johnmaeda]

GLOBAL MAXIMUM MINDSET: Sometimes, the smartest way to get better results from GenAI is to reset its context (literally) and give it a fresh start, not just keep iterating around a local maximum. This simple context engineering approach wasn’t obvious to me at first, but now I can see its brilliance. It’s a wonderful parable for life. Because starting over isn’t failure; it’s often the key to finding your true global maximum. https://t.co/rom1Q1epuf