Top all-time · #59
The latest research from Google
Company blog · research.google/blogaiscience
The latest research from Google is the company blog at research.google, writing about ai and science. Since 2020, 551 different Hacker News readers have linked it 757 times, most often “Zanzibar: Google’s Consistent, Global Authorization System (2019)”. It is still writing: the last post was on 2026-10-27, and you can follow it by feed.
Rarely recommended · 2 of 40 sampled comments recommend it?
Why HN links it
HN readers link Google Research's blog and papers on systems and computing, such as Borg, Bigtable, the monorepo, and the Willow quantum chip, along with AI research announcements.
Summarised by Claude from a sample of 40 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
Discover Google Research. We publish research papers across a wide range of domains and share our latest developments in AI and science research.
Most-linked posts?
- Zanzibar: Google’s Consistent, Global Authorization System (2019)19 people · 23 mentions · first linked 2021-02-24
- Nested Learning: A new ML paradigm for continual learning13 people · 14 mentions · first linked 2025-11-07
- Nearly all binary searches and mergesorts are broken (2006)9 people · 9 mentions · first linked 2024-07-13
- TurboQuant: Redefining AI efficiency with extreme compression9 people · 9 mentions · first linked 2026-03-24
- Machine Learning: The High Interest Credit Card of Technical Debt8 people · 13 mentions · first linked 2020-02-29
- Pipe Syntax in SQL8 people · 8 mentions · first linked 2024-08-22
- Google Titans architecture, helping AI have long-term memory8 people · 8 mentions · first linked 2025-12-04
- Snap: A microkernel approach to host networking (2019)8 people · 8 mentions · first linked 2020-02-18
- /pubs/pub43438/7 people · 8 mentions · first linked 2020-06-09
- Exploring a space-based, scalable AI infrastructure system design7 people · 7 mentions · first linked 2025-11-04
What Hacker News says?
Looking at the other comments, it seems highly likely zanzibar[1] is down. Their auth service. [1] PDF: https://research.google/pubs/pub48190.pdf
Comment linking /pubs/pub48190.pdf
on “Google outage – resolved”Submitted A return to hand-written notes by learning to read and write as a story · 595 points
HN discussionhttps://research.google/blog/validating-random-circuit-sampl...
Comment linking /blog/validating-random-circuit-sampling-as-a-benchmark-for-measuring-quantum-progress/
on “Willow, Our Quantum Chip”Submitted Google Titans architecture, helping AI have long-term memory as a story · 587 points
HN discussionIt's really important to note that the error correction test and the random circuit test are separate tests. The error correction is producing a single logical qubit of quantum memory, i.e. a single qubit with no gates applied to it. Meanwhile, the random circuit sampling uses physical qubits with no error correction, and is used as a good benchmark in part because it can prove "quantumness" even in the presence of…
Comment linking /blog/validating-random-circuit-sampling-as-a-benchmark-for-measuring-quantum-progress/
on “Willow, Our Quantum Chip”Submitted Accelerating scientific breakthroughs with an AI co-scientist as a story · 371 points
HN discussionEdit after skimming arxiv preprint[1]: Yeah, this is pretty huge. They achieved the result with surface codes, which are general ECCs. The repetition code was used to further probe quantum ECC floor. "Just POC" likely doesn't do it justice. (Original comment): Also quantum dabbler (coincidentally dabbled in bitflip quantum error correction research). Skimmed the post/research blog. I believe the key point is the sca…
Comment linking /blog/making-quantum-error-correction-work/
on “Willow, Our Quantum Chip”Submitted Learn Your Way: Reimagining Textbooks with Generative AI as a story · 359 points
HN discussionIt's in the article. Random circuit sampling benchmark: https://research.google/blog/validating-random-circuit-sampl...
Comment linking /blog/validating-random-circuit-sampling-as-a-benchmark-for-measuring-quantum-progress/
on “Willow, Our Quantum Chip”Submitted Android Earthquake Alerts: A global system for early warning as a story · 333 points
HN discussionThe slightly mind blowing bit is detailed here: > https://research.google/blog/making-quantum-error-correction... “the first quantum processor where error-corrected qubits get exponentially better as they get bigger” Achieving this turns the normal problem of scaling quantum computation upside down.
Comment linking /blog/making-quantum-error-correction-work/
on “Willow, Our Quantum Chip”Submitted Mathematical Optimization for Cargo Ships as a story · 324 points
HN discussionNo, back in the late 90s and early 00s, people were trying to engineer custom nucleases and transcription factors, my work was on doing molecular dynamics simulations to optimize TF sequence specificity (similar to engineered zinc fingers) for gene therapy. I wrote up my dissertation and published it in 2001, and then went off to find enough compute, IO, and smart people to make it happen ( https://research.google/b…
Comment linking /blog/groundbreaking-simulations-by-google-exacycle-visiting-faculty/
on “Sharing AI progress in mathematics”Submitted Pipe Syntax in SQL as a story · 308 points
HN discussionAccording to Google, Gboard uses Federating Learning to train a model on user data on the local device, so no sensitive data is not sent to the server. Only the gradients are sent and aggregated on the server. https://research.google/pubs/pub47586/
Comment linking /pubs/pub47586/
on “Grand jury subpoena for Signal user data, Central District of California”Submitted Satellite powered estimation of global solar potential as a story · 265 points
HN discussionI believe Macaroons[1] attempt to solve this problem. [1] https://research.google/pubs/pub41892/
Comment linking /pubs/pub41892/
on “Stealing Your Private YouTube Videos, One Frame at a Time”Submitted Sleep framework: A new way for LLMs to retain knowledge in continual learning as a story · 264 points
HN discussionThis blog post was mid-2020, so presumably a bit before that: https://research.google/blog/recent-advances-in-google-trans...
Comment linking /blog/recent-advances-in-google-translate/
on “A small number of samples can poison LLMs of any size”Submitted ScreenAI: A visual LLM for UI and visually-situated language understanding as a story · 262 points
HN discussionI'd love to see the research here. Because searching the software engineering Google papers from 2021-2021, the only somewhat relevant papers I found were case studies (such as [1] and [2]), which were based on Googlers' perception of their own processes. While self-perception can indeed by used to measure efficacy, there's a reason why it's used with great caution in surveys. I'll be honest, I have a hard time beli…
Comment linking /pubs/pub43835/
on “Google's infamous internal 2010 “I just want to serve 5TB” video now public”Submitted AI in software engineering at Google: Progress and the path ahead as a story · 253 points
HN discussionYep https://research.google/blog/accelerating-scientific-breakth...
Comment linking /blog/accelerating-scientific-breakthroughs-with-an-ai-co-scientist/
on “AlphaEvolve: A Gemini-powered coding agent for designing advanced algorithms”Submitted Zanzibar: Google’s Consistent, Global Authorization System (2019) as a story · 238 points
HN discussionCool, but don't get me wrong, isn't this essentially similar to Google's Co-Scientist, where multiple models are in a loop, passing context back and forth validating things? At its core, it's still a system of LLMs, which is impressive in execution but not fundamentally new. LLMs are undoubtedly useful at tasks like code "optimisation" and detecting patterns or redundancies that humans might overlook, but this annou…
Comment linking /blog/accelerating-scientific-breakthroughs-with-an-ai-co-scientist/
on “AlphaEvolve: A Gemini-powered coding agent for designing advanced algorithms”Submitted Scaling up linear programming with PDLP as a story · 221 points
HN discussionVery large, branchy programs with very short basic blocks. This describes all the major workloads at Google, according to their paper at https://research.google/pubs/pub48320/
Comment linking /pubs/pub48320/
on “Intel's disruption is now complete”Submitted Solving the minimum cut problem for undirected graphs as a story · 214 points
HN discussionMonorepos are pretty common nowadays, arguably popularized by Google https://research.google/pubs/pub45424/ That said (1) “you are not Google” as the saying goes (2) Google built their VCS from scratch in house for scale. However it’s not tops hard to do this with just git and some rules/patterns for organization.
Comment linking /pubs/pub45424/
on “Npm Audit: broken by design?”Submitted Transformers in music recommendation as a story · 211 points
HN discussion
