Hacker News: diwank

Breakthrough Prize Announces 2026 Laureates

diwank — Sun, 19 Apr 2026 05:41:28 +0000

Article URL: https://breakthroughprize.org/News/98

Comments URL: https://news.ycombinator.com/item?id=47822059

Points: 3

# Comments: 0

Reasoning models encode tool choices before they start reasoning

diwank — Sat, 04 Apr 2026 02:11:26 +0000

Article URL: https://arxiv.org/abs/2604.01202

Comments URL: https://news.ycombinator.com/item?id=47634936

Points: 3

# Comments: 0

New comment by diwank in "Show HN: Postgres extension for BM25 relevance-ranked full-text search"

diwank — Wed, 01 Apr 2026 02:20:48 +0000

had a bad experience with pg_search (paradedb) in the past

New comment by diwank in "Show HN: Postgres extension for BM25 relevance-ranked full-text search"

diwank — Wed, 01 Apr 2026 02:20:02 +0000

we have been using pg_textsearch in production for a few weeks now, and it's been fairly stable and super speedy. we used to use paradedb (aka pg_search -- it's quite annoying that the two or so similarly named), but paradedb was extremely unstable, led to serious data corruption a bunch of times. in fact, before switching to pg_textsearch, we just switched over to plain trigram search coz paradedb was tanking our db so often...

also shoutout to tj for being super responsive on github issues!

New comment by diwank in "Day 1 of ARC-AGI-3"

diwank — Fri, 27 Mar 2026 06:23:26 +0000

this is so disingenuous on symbolica's part. these insincere announcements just make it harder for genuine attempts and novel ideas

Show HN: Datetime-bench: which datetime formats LLMs get right (and wrong)

diwank — Thu, 26 Mar 2026 21:11:40 +0000

tl;dr

* If you need an LLM to parse OR emit a timestamp, use:

  RFC 3339 ( e.g. 2024-03-26 10:30:00-05:00 )

* python date format also works well

* Do NOT use unix epoch or javascript date formats.

* Smaller models and non-reasoning models still make a LOT of mistakes in time parsing / formatting.

---

There are lots of temporal reasoning benchmarks (like TimeBench, TRAM, etc.) but they test whether models understand time concepts. Nothing on which datetime output format models get right most often. So we just built the benchmark ourselves.

We tested 22 models across Google, Anthropic, OpenAI, Qwen, and GLM on 235 scenarios and 7 different formats.

The two that surprised us the most were JavaScript Date and unix epoch. JavaScript Date is probably the most commonly used format and it's wrong ~1 in 4 times on parsing. Unix epoch drops to 40% on arithmetic tasks. If you need epoch, just have the model output a string and convert it yourself in code.

Comments URL: https://news.ycombinator.com/item?id=47535806

Points: 2

# Comments: 1

New comment by diwank in "Antimatter has been transported for the first time"

diwank — Wed, 25 Mar 2026 18:53:41 +0000

Angels & Demons anyone?

Show HN: Dropbox for Your Agents' Memories

diwank — Wed, 25 Feb 2026 22:52:18 +0000

Article URL: https://memory.store/start

Comments URL: https://news.ycombinator.com/item?id=47159199

Points: 2

# Comments: 0

New comment by diwank in "I want to wash my car. The car wash is 50 meters away. Should I walk or drive?"

diwank — Mon, 16 Feb 2026 07:21:36 +0000

opus 4.6 gets it right more than half the times

Show HN: IQT – Why space feels panoramic and time feels fleeting

diwank — Thu, 12 Feb 2026 02:15:12 +0000

I've spent the past year building a theory of phenomenal experience (consciousness) that's designed to be falsifiable.

It identifies phenomenal quality with a specific mathematical object (the algebra-state pair on a causal diamond in algebraic quantum field theory) and derives three experimental protocols with pre-registered failure conditions.

The short version: take the "intrinsic nature" intuition from Russellian monism seriously, ask what mathematical object could play that role, and discover that constraint-based reasoning forces you to a unique answer. Then show that this answer dissolves the composition problem, generates temporal phenomenology from diamond geometry, and predicts specific measurable differences from IIT (overlapping conscious regions vs. exclusion) and GNW (experience without global broadcast).

The paper is ~35 pages, mathematically involved but with interpretive boxes and discipline-specific reading guides. I'm actively seeking collaborators in anesthesiology, epilepsy surgery, and psychedelic research to run the protocols.

Paper: https://github.com/creatorrr/intrinsic-quality-theory/blob/m...

I'd especially value pushback from mathematical physicists on the AQFT formalization and from neuroscientists on whether the effective-theory bridge actually constrains the empirical predictions or is doing post-hoc work. Tell me what breaks.

Comments URL: https://news.ycombinator.com/item?id=46984122

Points: 1

# Comments: 0

New comment by diwank in "Ask HN: What are you working on? (February 2026)"

diwank — Mon, 09 Feb 2026 03:02:38 +0000

Working on Memory Store: persistent, shared memory for all your AI agents.

https://memory.store

The problem: if you use multiple AI tools (Claude, ChatGPT, Cursor, etc.), none of them know what the others know. You end up maintaining .md files, pasting context between chats, and re-explaining your project every time you start a new conversation. Power users spend more time briefing their agents than doing actual work.

Memory Store is an MCP server that ingests context from your workplace tools (Slack, email, calendar) and makes it available to any MCP-compatible agent. Make a decision in one tool, the others know. Project status changes, every agent is up to date.

We ran 35 in-depth user interviews and surveyed 90 people before writing a line of product code — 95% had already built workarounds for this problem (custom GPTs, claude.md templates, copy-paste workflows). The pain is real and people are already investing effort to solve it badly.

Early users are telling us things like one founder who tracked investor conversations through Memory Store and estimated talking to 4-5x more people because his agents could draft contextual replies without manual briefing. It helped close his round.

Live in beta now. Would love feedback from anyone who's felt this pain! :)

New comment by diwank in "GPT-5.2 and GPT-5.2-Codex are now 40% faster"

diwank — Wed, 04 Feb 2026 00:56:07 +0000

I dont think this is Cerebras. Running on cerebras would change model behavior a bit and it could potentially get a ~10x speedup and it'd be more expensive. So most likely this is them writing new more optimized kernels for Blackwell series maybe?

Putting Gemini to Work in Chrome

diwank — Thu, 29 Jan 2026 03:43:52 +0000

Article URL: https://blog.google/products-and-platforms/products/chrome/gemini-3-auto-browse/

Comments URL: https://news.ycombinator.com/item?id=46805557

Points: 56

# Comments: 63

From kinetic gases to an exponentially expanding universe

diwank — Wed, 14 Jan 2026 05:29:07 +0000

Article URL: https://iopscience.iop.org/article/10.1088/1475-7516/2025/10/050

Comments URL: https://news.ycombinator.com/item?id=46612677

Points: 1

# Comments: 0

New comment by diwank in "A battle over Canada’s mystery brain disease"

diwank — Sun, 11 Jan 2026 06:55:10 +0000

yeah I agree. this is really unfortunate because it seems that there is something systemic here at play which has become twisted up in a cult of personality and that's made a rigorous scientific investigation very difficult

New comment by diwank in "OpenAI are quietly adopting skills, now available in ChatGPT and Codex CLI"

diwank — Fri, 19 Dec 2025 08:38:23 +0000

GitHub Copilot now supports Agent Skills

https://github.blog/changelog/2025-12-18-github-copilot-now-...

DeepSeek-v3.2 and v3.2 Speciale Announced

diwank — Mon, 01 Dec 2025 13:49:42 +0000

Article URL: https://huggingface.co/deepseek-ai/DeepSeek-V3.2/tree/main

Comments URL: https://news.ycombinator.com/item?id=46107357

Points: 5

# Comments: 0

Effective harnesses for long-running agents

diwank — Fri, 28 Nov 2025 19:05:16 +0000

Article URL: https://www.anthropic.com/engineering/effective-harnesses-for-long-running-agents

Comments URL: https://news.ycombinator.com/item?id=46081704

Points: 125

# Comments: 37

New comment by diwank in "Show HN: Why write code if the LLM can just do the thing? (web app experiment)"

diwank — Sat, 01 Nov 2025 21:57:40 +0000

Just in time UI is incredibly promising direction. I don't expect (in the near term) that entire apps would do this but many small parts of them would really benefit. For instance, website/app tours could be just generated atop the existing ui.

Cursor 2.0 – A step in the wrong direction

diwank — Fri, 31 Oct 2025 00:32:28 +0000

Article URL: https://forum.cursor.com/t/2-0-a-step-in-the-wrong-direction/139648

Comments URL: https://news.ycombinator.com/item?id=45767089

Points: 3

# Comments: 1