How do we make sure fairness for course project, if some student can't afford Fable/Sol?
Unsolicited hot take on colleges & AI (at least wrt to CS). The scope of material doesn’t need to change much. But we need to hold students to a higher standard instead of race to the bottom. AI doesn’t turn people into 10x or 100x engineers. The leverage is proportional to…
Show more
We're working at the forefront of AI theorem proving technology to tackle exciting software and hardware verification challenges. If you have experience with formal verification in general, or with interactive theorem provers, please consider joining us! Work with us to develop our Aristotle AI theorem prover to go beyond what can be done with SOTA models, finally democratizing formal verification and making it available to everyone. We're based in Palo Alto and London.
Show more
So many confused and incorrect answers to this. There is absolutely no reason you can't have a deterministic LLM. Also many people are confused by what "determinism" means in this context. Let's go over everything people are confused about 🧵
Show more
Here's a new post about durability and write-ahead logs. Write-ahead logs are used almost everywhere. But to build an intuition for why, it is helpful to imagine what you would do without a WAL. And to explore the meaning of durability.
Show more
The Garnet research paper has received the best paper award at VLDB 2026! The work will be presented in Boston on Thursday, Sep 3. Thanks to all our collaborators, users, partners, & OSS contributors!
Links:
Show more
Our group is part of two @VLDBconf papers on query opt. and execution:
1) Ultron, a system for building history-based optimization rules, deployed at @databricks
2) Global hash tables strike back!(?), an experimental analysis of fully concurrent group by aggregation. ⬇️
Show more
How many different databases does your company currently use in production?
I'm seeing some rather small companies with an impressive collection (multiple datalakes, analytics stores, OLTP stores).
I wonder if thats the "new normal".
Show more
> We just raised and we're hiring, come join us!
> Careers page: 0 job openings.
👏
(this happens so often!)
How does your startup get featured on The Consensus At a Glance? For starters, make sure you have an RSS feed!
Only 213 of 741 tracked companies today have an RSS feed we could find.
If you go to your company page on The Consensus and see the "From Their Blog" section you're all set.
Otherwise I don't have an RSS feed for you. DM and I'm happy to add one.
Show more
@krismicinski Most durable so far are fundamentals which is what was already being taught
I use a different approach: no show means you can’t attend any of the events I organize in the future.
The no show ratio is close to zero, people not attending send me an email before the event so I can send the invite to someone in the waiting list.
Show more
Six hours of my favorite melancholy jazz.
@_Felipe @kellabyte It’s a lot more work to build a minimum viable ARIES btree than an LSM, and so we have less of them. But the number of really good ideas to play with have been ever increasing: TSBT/MVBT version splitting, Bf-Tree, Bwe-Tree. So we feel overdue for btrees being used at all.
Show more
Really enjoying Isabel Wilkerson's The Warmth of Other Suns. Very well written.
@_Felipe @kellabyte I think LSM Trees have won largely because there’s no good btree library. LMDB is great but single writer. AWS showed BDB struggles. Cloudflare ran with TokyoCabinet and it’s not terrible but not great. There’s nothing of the level of polish and reliability as RocksDB.
Show more
If you've wanted to write for The Consensus but you don't have the time, or it seemed daunting, I'm looking for folks to write short briefs for weekly roundups.
Come write with me!
Show more
We have a new (professionally-designed) logo!
