Hacker News: _davide_

New comment by _davide_ in "WiCi: Connecting GPUs over WiFi"

_davide_ — Tue, 03 Nov 2026 11:31:14 +0000

It's a wifi wrapper for an RTX GPU, they built a company around this?

New comment by _davide_ in "Mass attempted replication of old findings of LLMs with new models"

_davide_ — Tue, 03 Nov 2026 09:16:25 +0000

Seems like this claim is expired in 3...2...1...

New comment by _davide_ in "GPT-6.1 Terra"

_davide_ — Tue, 27 Oct 2026 05:50:40 +0000

6.0 re branded as terra?

New comment by _davide_ in "Attack of the Vibe Coded Photoshop Clones"

_davide_ — Mon, 12 Oct 2026 07:29:27 +0000

I've been vibe coding plenty of apps for personal use, especially to replace free Google stuff; I'm doing it for lack of trust and control those apps give me. Even more would apply to Adobe stuff.

Innovation in Adobe products? Really? I mean, really? Are you serious?

And the main argument is performance on high resolution?

New comment by _davide_ in "Typesafe AI raises $870M at $7.5B"

_davide_ — Sat, 10 Oct 2026 07:35:45 +0000

I guess they are investing in the team rather than on their demo product

New comment by _davide_ in "Kagi Ends Orion Browser Development for Linux, Will Open-Source It"

_davide_ — Mon, 05 Oct 2026 06:24:46 +0000

Kde html -> webkit -> blink (which is the html of chrome). There is a huge work around the mere rendering engine, to the point that its way more common to fork the whole browser rather than just the low level engine.

New comment by _davide_ in "GPT-6.1 Sol replaces GPT-6 Sol after just 7 days, with near-Astra intelligence"

_davide_ — Wed, 30 Sep 2026 13:02:20 +0000

> it would make absolutely no sense to sit on it.

Yeah, it does, it might be misaligned, a snapshot of the going on training, bigger than they can serve publicly, not yet completed the full training pipeline.

If you see the knowledge cutoff you can see that sol 5.6 finished the initial main (+stage edited) of the training pipeline on Feb 16, 2026 but it was publicly released on July 9.

The opposite would be weird: if they do NOT have an unreleased in-house model that would be really odd.

New comment by _davide_ in "Dots: Always-on agents"

_davide_ — Wed, 30 Sep 2026 05:08:58 +0000

Claude code is a terribile harness, visibility is below zero, require a terminal session per each workspace, remote access makes you wonder if you should give copilot deluxe + a try. And according to the TOS you can't you whatever you want

New comment by _davide_ in "GPT 6.1 Sol: Near-Astra intelligence for a fifth of the price"

_davide_ — Tue, 29 Sep 2026 19:21:09 +0000

As a reference i burn 1% percent for every 40 minutes of sol on average

New comment by _davide_ in "OpenAI: Tomorrow we are re-opening the Pro $200 subscription"

_davide_ — Tue, 29 Sep 2026 09:00:58 +0000

Not going back to anthropic as i love the alignment of 5.6 sol, but I'll start looking around

New comment by _davide_ in "Show HN: The last day of the dinosaurs, as an interactive painting"

_davide_ — Fri, 25 Sep 2026 09:34:11 +0000

why

New comment by _davide_ in "Show HN: Whiteboard (YC W26) – An open-source IDE for thoughtful software design"

_davide_ — Fri, 25 Sep 2026 09:01:49 +0000

The UX seems nice, but the scope is way too narrow. I would be actually lazier for me to to just rebuild it inside my own harness (exactly as i want it) than start looking at yours.

New comment by _davide_ in "Is A.I. Above the Law?"

_davide_ — Thu, 24 Sep 2026 19:16:05 +0000

I'll quote myself:

> It's not AI that's ignoring the law! It's the OAI that's breaking IT!

> I hope every single journalist who tries to pin responsibility on an LLM gets 100 days of continuous painful diarrhea.

New comment by _davide_ in "OpenAI agent hacked Australian government website, PM says"

_davide_ — Thu, 24 Sep 2026 14:15:40 +0000

> but more important is that this is not the first instance of AI agents ignoring laws on accessing online information.

It's not AI that's ignoring the law! It's the OAI that's breaking IT!

I hope every single journalist who tries to pin responsibility on an LLM gets 100 days of continuous painful diarrhea.

New comment by _davide_ in "Jev in 25 Lines of Python"

_davide_ — Wed, 23 Sep 2026 10:45:55 +0000

Nice idea! Didn't think about that; a single linear memory allocation could do the trick

New comment by _davide_ in "Jev in 25 Lines of Python"

_davide_ — Wed, 23 Sep 2026 08:46:28 +0000

> <200ms for 45 questions at once

Considering your own question length: ~120 characters x 45 divided by 4.1 ~= 1317 tokens.

So question processing at 5.5k PP(around the actual PP speed of GPT5.6 Sol) it would take around ~0.24 seconds + the context processing.

Computing the output should be around ~20ms (at 50 tok/s), computing 45 tokens in parallel.

> have 0% malformed output

Pretty trivial; only the allowed output is selectable :)

So, I keep repeating myself: Jev was a low-hanging fruit all along; no one cared, and probably no one will in a few weeks?

New comment by _davide_ in "Jev in 25 Lines of Python"

_davide_ — Wed, 23 Sep 2026 08:30:39 +0000

By design it can't be significantly slower than Jev: the prompt processing (AKA PP) is exactly the same on both and will take most of the time. Then you can process every single "question" in parallel, just predicting one or two tokens (if an answer is ambiguous with a single token) per each question, again in a single batch.

So, fast in the LLM space and comparable with Jev.

New comment by _davide_ in "Jev in 25 Lines of Python"

_davide_ — Wed, 23 Sep 2026 08:25:31 +0000

Agreed, it's a real issue, but it can probably be vastly reduced by having the schema in the system prompt and by giving the model an expectation of a fixed value: no decent modern would pick a prose ligament over a provided value.

To completely squash the issue, a few cheap LoRa iterations will do the trick just fine.

New comment by _davide_ in "Introducing System One Models and Jev"

_davide_ — Wed, 16 Sep 2026 05:46:37 +0000

This is too much for me. ML playing doom was a thing since before LLMs, decisions tree were always insanely and no one ever used then anyway, i can't see anything new in this yet everyone is treating this as a revolution. This technology was always there and quite easily accessible all along.

New comment by _davide_ in "Introducing System One Models and Jev"

_davide_ — Tue, 15 Sep 2026 20:40:46 +0000

What's the difference compared to just taking an embedding and feed forward a simple net trained for the task?