they insisted on
- strict serializability by default, no compromises
- horizontal scaling across nodes while maintaining the above
- seamless replication and repair with instant failover
- extremely stringent correctness testing
- strict serializability by default, no compromises
- horizontal scaling across nodes while maintaining the above
- seamless replication and repair with instant failover
- extremely stringent correctness testing
November 2, 2026 at 7:14 PM
they insisted on
- strict serializability by default, no compromises
- horizontal scaling across nodes while maintaining the above
- seamless replication and repair with instant failover
- extremely stringent correctness testing
- strict serializability by default, no compromises
- horizontal scaling across nodes while maintaining the above
- seamless replication and repair with instant failover
- extremely stringent correctness testing
Here is a mass attempted replication of old findings of LLMs with new models. Much conventional wisdom is outdated.
Recent models no longer invent citations and both hallucination rates & stereotyping has dropped. But they are still sycophantic & lean left. zou-group.github.io/llm-claims-s...
Recent models no longer invent citations and both hallucination rates & stereotyping has dropped. But they are still sycophantic & lean left. zou-group.github.io/llm-claims-s...
November 3, 2026 at 2:38 AM
Here is a mass attempted replication of old findings of LLMs with new models. Much conventional wisdom is outdated.
Recent models no longer invent citations and both hallucination rates & stereotyping has dropped. But they are still sycophantic & lean left. zou-group.github.io/llm-claims-s...
Recent models no longer invent citations and both hallucination rates & stereotyping has dropped. But they are still sycophantic & lean left. zou-group.github.io/llm-claims-s...
But genes are NOT replicators. They get replicated. Replication is not something they DO. They're not agents and they're not alive.
October 31, 2026 at 4:42 PM
But genes are NOT replicators. They get replicated. Replication is not something they DO. They're not agents and they're not alive.
Also separate from the lobbying question is the much murkier and more difficult to interrogate issue of genuine ideological replication— how effective would the pro Israel lobby really be if there was not a genuine well of historic identification amongst DC types with Israel?
November 3, 2026 at 12:04 AM
Also separate from the lobbying question is the much murkier and more difficult to interrogate issue of genuine ideological replication— how effective would the pro Israel lobby really be if there was not a genuine well of historic identification amongst DC types with Israel?
GTA6 truly is a child's idea of what a cool game is, utterly fascinated by the replication of real life with absolutely no care for whether the reality simulacrum is a good or enjoyable thing. No concern for how these systems fit together or what the experience of engaging with them is like.
November 2, 2026 at 7:30 PM
GTA6 truly is a child's idea of what a cool game is, utterly fascinated by the replication of real life with absolutely no care for whether the reality simulacrum is a good or enjoyable thing. No concern for how these systems fit together or what the experience of engaging with them is like.
Many analyst large scale replication project when?
November 3, 2026 at 8:47 AM
Many analyst large scale replication project when?
I actually think Azmi Bishara describes this process of ideological replication and convergence really well in his 2021 book
October 30, 2026 at 4:44 PM
I actually think Azmi Bishara describes this process of ideological replication and convergence really well in his 2021 book
Under a metabolism-first view of life (as opposed to the replication-first view favored by Dawkins), genes came late to the party.
October 31, 2026 at 4:44 PM
Under a metabolism-first view of life (as opposed to the replication-first view favored by Dawkins), genes came late to the party.
Conservatism = replication (of what never was)
October 30, 2026 at 3:29 PM
Conservatism = replication (of what never was)
🙏🙏🙏 hoping for great replication success on my side
November 4, 2026 at 2:28 PM
🙏🙏🙏 hoping for great replication success on my side
We are hosting Replication Games on January 21, 2027 at University of Cologne. Economic articles will be reproduced! Register here:
www.surveymonkey.ca/r/cologne_re...
www.surveymonkey.ca/r/cologne_re...
November 3, 2026 at 7:48 PM
We are hosting Replication Games on January 21, 2027 at University of Cologne. Economic articles will be reproduced! Register here:
www.surveymonkey.ca/r/cologne_re...
www.surveymonkey.ca/r/cologne_re...
Good news everyone, Thundersnap 0.04 is out and it's also finally mirrored on tangled. tangled.org/apenwarr.ca/...
I've been using this as my daily main linux environment for several months and it's getting quite pleasant.
I've been using this as my daily main linux environment for several months and it's getting quite pleasant.
thundersnap by apenwarr.ca · Tangled
A Linux container environment where everything is mutable, undo is instant, and cross-machine replication is trivial.
tangled.org
November 3, 2026 at 6:25 PM
Good news everyone, Thundersnap 0.04 is out and it's also finally mirrored on tangled. tangled.org/apenwarr.ca/...
I've been using this as my daily main linux environment for several months and it's getting quite pleasant.
I've been using this as my daily main linux environment for several months and it's getting quite pleasant.
Halloween is a day to play!
Happy Halloween. Be safe & sane & sassy & silly...but most of all, shiny.
Happy Halloween. Be safe & sane & sassy & silly...but most of all, shiny.
October 31, 2026 at 3:29 PM
Halloween is a day to play!
Happy Halloween. Be safe & sane & sassy & silly...but most of all, shiny.
Happy Halloween. Be safe & sane & sassy & silly...but most of all, shiny.
Alas the Stephen Miller replication academy where they teach young republicans how to shower came too late to save them
November 4, 2026 at 3:08 AM
Alas the Stephen Miller replication academy where they teach young republicans how to shower came too late to save them
Surprising non-transcriptional function for mediator kinase CDK8 in human DNA replication:
Corrado Santocanale & coworkers find it to cooperate with CDC7 in MCM helicase phosphorylation & activation–with implications for antiproliferative actions of CDC7 inhibitors
link.springer.com/article/10.1...
Corrado Santocanale & coworkers find it to cooperate with CDC7 in MCM helicase phosphorylation & activation–with implications for antiproliferative actions of CDC7 inhibitors
link.springer.com/article/10.1...
CDC7 and CDK8 kinases cooperate to support DNA replication origin firing in human cells - The EMBO Journal
The coordinated activation of DNA replication origins is important for efficient DNA synthesis and genome stability. S-phase cyclin-dependent kinases (CDKs), together with CDC7 kinase, are essential f...
link.springer.com
November 2, 2026 at 8:24 PM
Surprising non-transcriptional function for mediator kinase CDK8 in human DNA replication:
Corrado Santocanale & coworkers find it to cooperate with CDC7 in MCM helicase phosphorylation & activation–with implications for antiproliferative actions of CDC7 inhibitors
link.springer.com/article/10.1...
Corrado Santocanale & coworkers find it to cooperate with CDC7 in MCM helicase phosphorylation & activation–with implications for antiproliferative actions of CDC7 inhibitors
link.springer.com/article/10.1...
These have to do with Dawkins conception of life. He sees genes as replicators and life as replication of genes. All his arguments follow from this conception.
October 31, 2026 at 4:41 PM
These have to do with Dawkins conception of life. He sees genes as replicators and life as replication of genes. All his arguments follow from this conception.
The canonical "three-phase commit" system is GitHub itself: github.blog/engineering/...
Stretching Spokes
GitHub’s Spokes system stores multiple distributed copies of Git repositories. This article discusses how we got Spokes replication to span widely separated datacenters. Background: Spokes GitHub deve...
github.blog
November 3, 2026 at 5:23 PM
The canonical "three-phase commit" system is GitHub itself: github.blog/engineering/...
its the same "we have ice cream at home" energy... EVEN if its EXACTLY the same kind...
but looking back at being a kid I understand lol... it's the "experience" NOT the replication of the product lol... no matter how close you replicated it. lol...
but looking back at being a kid I understand lol... it's the "experience" NOT the replication of the product lol... no matter how close you replicated it. lol...
November 3, 2026 at 11:24 PM
its the same "we have ice cream at home" energy... EVEN if its EXACTLY the same kind...
but looking back at being a kid I understand lol... it's the "experience" NOT the replication of the product lol... no matter how close you replicated it. lol...
but looking back at being a kid I understand lol... it's the "experience" NOT the replication of the product lol... no matter how close you replicated it. lol...
Less than ideal wire protocol that uses a ton of memory and sends a ton of extra data over the wire on every transaction, no good active/active replication story, terrible behavior when you fill replication slots, indexes are tied to glibc versions, the db driver in go is non standard, etc...
November 2, 2026 at 9:19 PM
Less than ideal wire protocol that uses a ton of memory and sends a ton of extra data over the wire on every transaction, no good active/active replication story, terrible behavior when you fill replication slots, indexes are tied to glibc versions, the db driver in go is non standard, etc...
“The replication crisis, if there is one, cannot be established by methods used to declare it” (Buzbas & Devezer, 2026) – and that claims about rigour must themselves meet rigorous standards (Bak-Coleman & Devezer, 2024)."
📣 The difference between replicable and not replicable is not itself scientifically replicable. 📣
New work with Erkan Buzbas, showing that verdicts such as "X% of results replicated" are based on an inferential machinery that doesn't work.
arxiv.org/abs/2604.26268
New work with Erkan Buzbas, showing that verdicts such as "X% of results replicated" are based on an inferential machinery that doesn't work.
arxiv.org/abs/2604.26268
November 3, 2026 at 9:40 AM
“The replication crisis, if there is one, cannot be established by methods used to declare it” (Buzbas & Devezer, 2026) – and that claims about rigour must themselves meet rigorous standards (Bak-Coleman & Devezer, 2024)."
Camerer also ran the 2016 replication of 18 lab experiments in economics: 11 held up, with replication effect sizes about two-thirds of the originals. Did that come up, and does he think neuroeconomics, with its small fMRI samples, has had its own replication pass yet?
November 2, 2026 at 3:03 PM
Camerer also ran the 2016 replication of 18 lab experiments in economics: 11 held up, with replication effect sizes about two-thirds of the originals. Did that come up, and does he think neuroeconomics, with its small fMRI samples, has had its own replication pass yet?
No scale? No problem.
But off the top of my head I've run into: terrible concurrency w/per process connections, garbage and VACUUM overhead/performance, crappy heuristics query planner, WAL replication but DIY failover, no horizontal write scaling, write contention, jittery query latency, etc.
But off the top of my head I've run into: terrible concurrency w/per process connections, garbage and VACUUM overhead/performance, crappy heuristics query planner, WAL replication but DIY failover, no horizontal write scaling, write contention, jittery query latency, etc.
there is no reason to make a new database. postgres can do it. if you think postgres can't do it, you just haven't found the right way to enable it in pg_settings yet
November 2, 2026 at 7:39 PM
No scale? No problem.
But off the top of my head I've run into: terrible concurrency w/per process connections, garbage and VACUUM overhead/performance, crappy heuristics query planner, WAL replication but DIY failover, no horizontal write scaling, write contention, jittery query latency, etc.
But off the top of my head I've run into: terrible concurrency w/per process connections, garbage and VACUUM overhead/performance, crappy heuristics query planner, WAL replication but DIY failover, no horizontal write scaling, write contention, jittery query latency, etc.
I just published SAP Data Integration: Federation vs. Replication Explained medium.com/@socialelina...
SAP Data Integration: Federation vs. Replication Explained
Should you query data where it lives or copy it into another environment? This is a common decision in SAP data integration and an…
medium.com
November 2, 2026 at 12:59 PM
I just published SAP Data Integration: Federation vs. Replication Explained medium.com/@socialelina...
You can now plug thundersnap in as an MCP server for any agent harness (aperture, pi, etc) and it provides a full shell environment that's not your local machine. Mutable thundersnap frames, replication, services, etc, even if you don't want to give it access to your local machine.
November 3, 2026 at 6:25 PM
You can now plug thundersnap in as an MCP server for any agent harness (aperture, pi, etc) and it provides a full shell environment that's not your local machine. Mutable thundersnap frames, replication, services, etc, even if you don't want to give it access to your local machine.