Romain

Homo sapiens. My shoe size is a European 45.5 and I prefer my orange juice with bits. CEO at @Thingiverse, former founding CEO @MyMiniFactory

[@levelsio @levelsio]

@airthings request: I'd like to pull in the air sensor history data from my Airthings sensors back to when I bought them, but the API is limited to 1000 I think, here's what my Claude Code says to you:

"Hi,

I'm building a personal integration against the Consumer API for my View Plus devices (using the client_credentials flow, scope read:device:current_values). Auth, GET /v1/devices, and GET /v1/devices/{id}/latest-samples all work great.

The issue is retrieving historical data. GET /v1/devices/{id}/samples returns 200 with the expected parallel arrays (time, co2, temp, humidity, …), but it only ever returns the first ~1,000 samples from the device's start date (for me, 2026-04-30 → 2026-05-02), and I can't get anything newer or page past it.

Things I've tried, all returning that same first 1,000-sample block:
- ?from=<epoch>&to=<epoch> with epoch seconds — from/to appear to be ignored (a from = now − 3 days window still returns April data).
- Same with millisecond timestamps.
- No params at all.
- offset, page, sn, skip — no effect.

Questions:
1. Is there a supported way to paginate through the full history, or to fetch a specific time window? If so, what are the correct parameter names/units?
2. Is the ~1,000-sample response a hard cap, and if so how do I request the next page?
3. If the Consumer API doesn't support historical/bulk retrieval, is that only available on the Business API / a different plan? I'd like to pull the full history for my own devices (a few months of data)."

[Thingiverse @thingiverse]

(1/4) Thingiverse is joining the MyMiniFactory family!

MyMiniFactory has acquired 100% of Thingiverse, bringing together two of the biggest communities in 3D printing.

[Tuo Liu @Robo_Tuo]

Ok everyone, here’s the European Humanoid Map.

Europe has more humanoids than I expected: Paris, France has four, Germany and the UK have two each, and Norway, Sweden, Luxembourg, Italy, and Spain each have one. More will emerge.

Exciting times ahead for sure. Cheers.

[Romain @RomainKidd]

@chamath Chart 2 shows there are no moats in coding agents. Devs switch to the best model/product. This is consistent with the release of research lab models (one month OpenAI has the best model for coding, the following month Anthropic has the best model, then Google steps up, etc.)

[Romain @RomainKidd]

@sidahuj I like @Stanford's STORM:

https://t.co/zIR8en6qnh

[Romain @RomainKidd]

@TaiwanSpecial Taiwan does show up as a destination though

[Romain @RomainKidd]

@HyperSalesman DM

[Romain @RomainKidd]

@OfficialLoganK @SmokeAwayyy Please have a better mobile app too 🙏

[Romain @RomainKidd]

@Jason AI is splitting: consumer models racing to free ($20/month -&gt; $0) vs pro models charging $$$ (e.g., Devin, o1 Pro). UI wins on the low end—OpenAI's model is good enough for 99%+ of my basic prompts, but the mobile app is (still) the best. Will some labs quite consumer altogether?

[Romain @RomainKidd]

@elonmusk @SmokeAwayyy The Italian bit rate is 2x higher when you factor in hand gestures.

[Romain @RomainKidd]

@galonaranjo_ @dannypostma @dannypostma Would love to find out as well, please share more

[Romain @RomainKidd]

@Kegaman_TF2 That one was in the afternoon, this one had a higher 6.7 magnitude, and happened off of Yilan: https://t.co/Aarv743y8Z

[Romain @RomainKidd]

@ranjanxroy @ReadMargins The food delivery business only makes sense with automation, allowing for close to 0 marginal cost deliveries. The current version of food delivery companies with human drivers will be remembered in history like the Yahoos of the world of Internet search (early manual iterations)

[Romain @RomainKidd]

@neonemesis Par contre,je ne pense pas que Netflix puisse réduire le débit beaucoup plus,sinon ils vont perdre des clients. Je les vois mal lancer une saison entière d'un nouvel épisode en même temps par exemple.Peut-être qu'ils vont se mettre à sortir des épisodes ttes les semaines

[Romain @RomainKidd]

@neonemesis Et ça montre aussi que la plupart des gens n'ont pas la moindre idée du fonctionnement d'une boîte internet...

[Romain @RomainKidd]

@RBReich The question is not "should we bail out corporations", the question is "should we bail out corporations so that people have a job (to get back to when things eventually ramp up again)". A simple rule could be that bail out money can only be used to preserve employment

[Romain @RomainKidd]

@mattturck Twitter makes money from advertising. Usage has probably gone up with people online at home all day, but my guess is that advertisers are slashing their budgets. So cost increase and revenue decrease ==&gt; Social is NOT a good BUSINESS right now

[Romain @RomainKidd]

The backstory of the Venus de Milo, one of the more popular 3D printable models on @MyMiniFactory @Scan_The_World. Please don't use red markers on the real thing!

[Phil Edwards @PhilEdwardsInc]

How did the Venus de Milo get so famous? https://t.co/45HHQH8LQI