ltrg

Joined 4 years ago

Karma
281
Submissions
33
Member since
Jan 29
I'm Head of Research at datadesk.eco, where I use commercial and open source data to model and investigate the global fossil fuel industry.

[email protected]

View on Hacker News

Recent submissions

Comment · 1 month ago
I'm not sure, to be honest -- I think the main agent loop, tools, etc. are fairly standard and the main draw is fast start-up time, but it does have a very minimal default system prompt too.
Comment · 1 month ago
Comment · 2 months ago
Or uses swipe gestures on mobile keyboard!
Comment · 2 months ago
I use research agents to attribute methane emissions plumes detected by satellites to oil and gas infrastructure on the ground, using a pre-baked database of geospatial data and web research.

Had a tool that called out from DeepSeek to Gemini 3.5 Flash for viewing the spatial features in the context of high-resolution satellite imagery of each site, but will be trialling this model for the whole thing now.

Comment · 2 months ago
Yes! Also everything's a "gate" that needs to be "wired up".
Comment · 3 months ago
Got tired of wiring up those gates
Comment · 4 months ago
It would be super interesting to see if HN mentions serve as a leading indicator of company performance/valuations -- I wouldn't be surprised.
Comment · 4 months ago
Not sure the parenthesis was required but I laughed.
Comment · 4 months ago
> This vaguely-nationalist world view around tech that’s emerging in Europe is dangerous, man.

It’s a response to the actually-nationalist practice of the United States. I can understand why it might feel different from California, but things are a bit scary over here right now.

Comment · 4 months ago
Fable seems very good at finding bugs (unsurprising given Mythos lineage), so this seems a pretty smart strategy. Once you see the bugs it finds in your existing Opus code, it's going to be hard to go back, psychologically speaking.
Comment · 5 months ago
Really cool stuff. Nitpick: it failed to grab an OSM ID for my house and fell back to postcode centroid, but then still reported LIDAR-derived shading at quite high precision.

I'm wondering if it should fall back to a more general shading approach when no OSM building footprint is available, to avoid false precision? My street has a gap in the houses on the other side from mine, so picking the right location matters for the calculation.

You could also try Inspire Index polygons instead of OSM? These correspond to actual lease/freehold boundaries.

Comment · 5 months ago
Comment · 5 months ago
"foundation model" a somewhat confusing phrase in the title here, given the majority of HN front page.
Comment · 6 months ago
Just noticed this in my inbox from earlier. Seems to only affect a subset of feature-flagged users.

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton