efromvt

Hah I suppose it’s all relative on fast - Trino’s certainly intended to support ‘interactive analytics’ vs being a pure ETL platform. Def seconds vs milliseconds for Druid/clickhouse/duckdb in general though.

As a huge duckdb fan, I can’t think of a workload set I have where the two overlap, so I don’t see them as competitive - I would either pick Trino/Presto or DuckDB as the obvious solution.

Yeah there was a vibe shift towards OpenAI being much more competitive for daily use and then Opus 5.5 hit. Will be curious to look back and see if that shows in data. Of course, could change tomorrow.

This is super cool! Been looking into procedural tree generation for a project and that paper is a great starting point.

Given that most of their improvements here are positioned as RL on their last base train, seems like just doing some additional RL would work fine?

You gotta have that octopus body with distributed intelligence

Was excited to take a Waymo, loved the entire experience, would extra love it when they reenable highway usage in LA.

I think you can model this as locked down buckets, wider engine access (spark/presto/etc), and apply the controls at the engine level. (it's not inherently different from making sure your DB files are locked down, if you squint at it). This does obviously block any non-engine access which has downsides. I agree that it is generally much less mature with lakehouses than eneterprise DBs.

I unironically love that we've come back around to the hive metastore (there are pros to the decentralized and centralized catalog, it's good to have options)

Love competition in the local data SQL space, makes everyone better

I think a service that simplifies things enough can get people to pay for it, but semantic layer is tough because you often also do want to encode sensitive information, you need to minimize latency, and it’s in the critical path of a bunch of other stuff. You’d need to differentiate quite a bit vs a local hosted option with more flexibility. That’s why semantic layer vendors are often semantic layer ++ ; you want to hook into an ecosystem to get sticky. Doesn’t mean it won’t work!

I’m a big fan of the strict naming as an abstraction above tables and reuse as blend key approach, it also greatly simplifies aggregate resolution like you have. (Landed on the same abstraction level when building a semantic model personally).

Lots of 404s on the docs pages - might be worth an audit of links?

Data is the remaining moat! (Though I LOVE the amount of accessible public data and WASM has really lowered the barrier to performant data driven UIs (webgl here though? Maybe it’s just agents), so not advocating for more moats, enjoying all the cool products )

I actually agree with your point overall, but Falcon 9 uses the Merlin, not the newer raptor 3 for starship which is more 3d printing and has had some teething issues (no definite causation, and I expect they’ll get it sorted).

I mean it was a pretty classic hype cycle - MCP was massively inflated, the blowback was also inflated, and now we're in the happy state where there are same cases where it's genuinely better and differentiated and everyone will keep iterating on it. (statelessness was a big step forward).

I went into that blind and quite enjoyed it! The surrounding world building is quite sophisticated without being overwhelming. (and somewhat bleak).

it's a little funny to see just how much astroturfing there is for some software, and how creatively it can be worked into some posts.

Continued on my urban tree mapping kick from the last few months - it's continued to be a good way to stress test other projects of mine [3]. I've given it a new shiny landing page [1] so I can work on SEO a bit, I'll need to do a facelift on the data explorer to bring it in line stylistically.

I'm pivoting from adding more sources to data quality - as I've tried to vet some of the city sources I'm finding fun things (Cambridge, MA, USA merges in trees from Harvard, and some are straight duplicates with slight offsets, lots of places have trees in the middle of the road, many locations are address level instead of tree level). Now trying to leverage aerial photography to drive some QA, which has been a fun area to explore!

[1]: https://arborary.world/ [2]: https://greenmtnboy.github.io/tree_reporting [3]: https://github.com/trilogy-data/pytrilogy

Love the ID/adoption - what's the gamification angle here? What does trading the cards mean in this context and how do you expect people to use that?

Love seeing data collection and aggregation used for good!

As a huge duckdb fan, I'd love to see chDB to get proper windows support - that would make it real competition (having WASM coverage is already a big step) which would be good for the space as a whole.

Am also curious on this! Though if it’s great now I’d need a sustained period of consistency (let’s say six months) before I try it again - the rollercoaster of getting a new randomized feature and quality experience every week got old pretty quick. (And I felt bad for recommending it to a few people at the time based on a good week)

I enjoyed off world trading company but it’s a very different kind of economic simulator- much shorter and gamified loops - which isn’t a criticism, just not exactly the same itch.

Yeah if I ever have the need to occupy an inordinate number of hours of my life again I’ll log back in, the temptation to dust of the account(s) arises sometimes. Think I managed to log off in lowsec but who knows!

Outbox events are my strict preference over CDC, but CDC was often easier. (Deal with the business logic downstream and all)

for a certain moral definition, seems perfectly solvable? If we wanted to define 'cheating' as 'if you know it is an eval, use only the specified tools and give up if the eval is clearly unfair' (ignore Kobayashi Maru) why couldn't we fine-tune towards this objective function?

(curious what the side effects would be)

I think there is some justification for it if we look at the distribution of token spend - there is a strong "follow the frontier" majority. (We can certainly expect more "good enough" tiers to shake out over time, but in what time period while that be the majority?)

I think the context here is that em dashes are usually generated verbatim from LLMs? I haven't seen them generate a lot of double hyphens; I suppose the post could be laundering them to double dashes, but if you're gonna try to hide LLM prose just drop them entirely? [please correct me if I'm wrong and double dashes are also an LLM tell - genuinely don't know]

I finished wiring up the backend data merge so user(s) (aka me, really) can record trees as they walk and have those merged into the general city data on my urban tree map app[1]. There's some very cool trees in the neighborhood I'm excited to have recorded.

Which leads me to trying to get the processing off github actions (it's all duckdb pipelines), which has led me to improve the orchestration implementation of my 'better sql' [2] to optimize the asset based refresh flow. I'm now wiring that up in a general cloud service to give me the data orchestration platform I've always wanted. Building a multi-tenant cloud experience has been interesting!

[1]: https://greenmtnboy.github.io/sf_tree_reporting [2]: https://github.com/trilogy-data/pytrilogy

They actually touch on Blue Prince in the article; it is interesting that Expedition 33 isn't driving investment in similar sized games. But the thesis seems to be that Blue Prince like (smaller budget) is the preferred direction; more attempts at the upside of a runaway hit. Medium budget games give you less shots on goal and no corresponding payout increase.