Hacker News: kevinventullo

New comment by kevinventullo in "US High school students' scores fall in reading and math"

kevinventullo — Wed, 10 Sep 2025 04:01:48 +0000

I thought charter schools and public schools received the same $/student.

New comment by kevinventullo in "I'm absolutely right"

kevinventullo — Fri, 05 Sep 2025 17:28:58 +0000

What’s funny is the jerky animation actually communicates so much more than the smooth animation.

New comment by kevinventullo in "A high schooler writes about AI tools in the classroom"

kevinventullo — Thu, 04 Sep 2025 23:23:40 +0000

I attended high school in the US in the early 00’s and cell phones were absolutely banned from classrooms. You could keep them in your locker and use them between classes, but that was it.

I attended college in the late 00’s, and I don’t think I took a single digital exam. Quizzes, sure, but for final exams even CS was pencil and paper (or a final project, which admittedly will have issues in the post-LLM era).

New comment by kevinventullo in "Hollow Knight: Silksong causes server chaos on Xbox, Steam, and Nintendo"

kevinventullo — Thu, 04 Sep 2025 18:37:23 +0000

FWIW there is a new-ish kind of intermediate genre between classic LAN/ranked multiplayer and single player, which is the whole “survival” genre. Generally speaking, they can be played as single player games, but also allow for small-scale co-op, synchronously or asynchronously. So even if you and a buddy have different schedules, you can make progress separately but still occasionally play together.

Valheim, Grounded, Ark, Satisfactory are a few among many others.

New comment by kevinventullo in "IBM and AMD to work on quantum-centric supercomputing"

kevinventullo — Tue, 26 Aug 2025 20:06:18 +0000

Basically anything that requires a massively parallel computation on undeterminable states that are only clear in hindsight.

From https://scottaaronson.blog/ :

“If you take nothing else from this blog: quantum computers won't solve hard problems instantly by just trying all solutions in parallel.”

New comment by kevinventullo in "The Size of Adobe Reader Installers Through the Years"

kevinventullo — Mon, 25 Aug 2025 14:39:01 +0000

What’s the best choice of free desktop PDF Viewer/Editor these days (any OS)?

On Windows I’ve been using PDF-XChange for a decade or so now, but curious if better alternatives have cropped up.

New comment by kevinventullo in "4chan will refuse to pay daily online safety fines, lawyer tells BBC"

kevinventullo — Fri, 22 Aug 2025 20:11:55 +0000

“Don't blame me, I voted for Kodos”

New comment by kevinventullo in "Project to formalise a proof of Fermat’s Last Theorem in the Lean theorem prover"

kevinventullo — Thu, 21 Aug 2025 20:08:32 +0000

That’s a fair point. But it greatly limits the scope of human-introduced error. I think already for FLT, the surface area for error in the kernel and in axiom translation is substantially smaller than the entirety of the literature which Wiles’s proof recursively depends on.

New comment by kevinventullo in "Sixteen bottles of wine riddle"

kevinventullo — Thu, 21 Aug 2025 17:16:31 +0000

You need at least 45 guesses since 2^44 < 16!

New comment by kevinventullo in "Project to formalise a proof of Fermat’s Last Theorem in the Lean theorem prover"

kevinventullo — Thu, 21 Aug 2025 04:39:48 +0000

I contend it is the only way to move forward on the goal of “automating” mathematics. Although we’ve seen natural language approaches do well at IMO, the human effort required to verify higher level proofs is too great with hallucinations being what they are. With something like Lean, you don’t need a human verifier.

New comment by kevinventullo in "DINOv3"

kevinventullo — Fri, 15 Aug 2025 18:05:54 +0000

Caveat: I am not an expert, so this is a semi-educated guess.

I imagine it would depend on whether DINOv3 captures the information of whether a given person is in the image, which I think is really a question about training data. So naively, I would guess the answer is yes for celebrities and no for non-celebrities. Partially for data/technical reasons, but also maybe due to the murkier legal expectation of privacy for famous people.

New comment by kevinventullo in "DINOv3"

kevinventullo — Thu, 14 Aug 2025 21:58:46 +0000

To elaborate, this is a foundation model. This basically means it can take an arbitrary image and map it to a high dimensional space H in which ~arbitrary characteristics become much easier to solve for.

For example (and this might be oversimplifying a bit, computer vision people please correct me if I’m wrong) if you’re interested in knowing whether or not the image contains a cat, then maybe there is some hyperplane P in H for which images on one side of P do not contain a cat, and images on the other side do contain a cat. And so solving for “Does this image contain a cat?”becomes a much easier problem, all you have to do is figure out what P is. Once you do that, you can pass your image into DINO, dot product with the equation for P, and check whether the answer is negative or positive. The point is that finding P is much easier than training your own computer vision model from scratch.

New comment by kevinventullo in "Gemma 3 270M: Compact model for hyper-efficient AI"

kevinventullo — Thu, 14 Aug 2025 18:44:30 +0000

Perhaps you can do some pre-processing before the LLM sees it, e.g. replacing every instance of “kill” with “NorwegianDudeGameKill”, and providing the specific context of what the word “NorwegianDudeGameKill” means in your game.

Of course, it would be better for the LLM to pick up the context automatically, but given what some sibling comments have noted about the PR risks associated with that, you might be waiting a while.

New comment by kevinventullo in "What's the strongest AI model you can train on a laptop in five minutes?"

kevinventullo — Thu, 14 Aug 2025 18:27:13 +0000

As I understand it, the most effective small models are synthesized from larger models.

New comment by kevinventullo in "Ask HN: How do you tune your personality to get better at interviews?"

kevinventullo — Thu, 14 Aug 2025 15:28:40 +0000

The rejection might also be random and out of your control. E.g. if they already filled the position or the higher-ups decided to save costs by taking back open headcount. Frustrating for sure, but not actionable.

From my POV, if you don’t have any strong signals about why you were rejected, I would just move on rather than trying to infer the reason.

New comment by kevinventullo in "Man develops rare condition after ChatGPT query over stopping eating salt"

kevinventullo — Wed, 13 Aug 2025 19:00:23 +0000

Is anyone else getting tired of these articles?

“Area man who had poor judgement ten years ago now has both poor judgement and access to chatbots”

New comment by kevinventullo in "The Anti-Pattern Game"

kevinventullo — Wed, 13 Aug 2025 06:18:14 +0000

In some sense, the program itself is a ~512 byte compression of an infinite stream of bytes.

New comment by kevinventullo in "How AI conquered the US economy: A visual FAQ"

kevinventullo — Fri, 08 Aug 2025 05:26:11 +0000

The benefits have just not been that wide ranging to the average person

I think you might be underestimating how many non-technical people are using LLM’s daily.

New comment by kevinventullo in "How AI conquered the US economy: A visual FAQ"

kevinventullo — Fri, 08 Aug 2025 05:14:52 +0000

Won’t you get margin called if the stock goes up in the meantime?

New comment by kevinventullo in "New executive order puts all grants under political control"

kevinventullo — Fri, 08 Aug 2025 03:36:14 +0000

Is anything stopping private philanthropy from funding science today? Generally speaking it is in the government’s interest to support fundamental scientific research.