One argument may be that ownership is the last role for a human in a business. The firm exists to show ownership of an AI and provides a mechanism for managing its proceeds.

Derek Muller of Veritasium has a presentation on why the promises of a revolution in education are never fulfilled. [0]

His hottest take is we're already close to the optimal process for learning, so technology isn't going to improve it. Learning takes work, and no technology can do the work for you.

[0] https://www.youtube.com/watch?v=0xS68sl2D70

Nanoclaw uses the official claude agent sdk, so it should be unaffected.

For those looking for an MIT alternative, there's an embeddable solution which uses PDFium (Apache) compiled to wasm instead of MuPDF (AGPL): https://www.embedpdf.com/

There's a hosted version for quick edits: https://app.embedpdf.com/

Discussion from several months ago: https://news.ycombinator.com/item?id=44901683

Neither fully handles XFA, but that's a perennial struggle.

> A recent study from Stanford

Source?

A recent study from Stanford

Even without prior knowledge, seeing others safely navigate the same section will lower your estimated risk.

> Musk miscalculated on 1) cost reduction in LIDAR

Given that Musk has a history of driving lower costs, it's unlikely he overestimated the long-term cost floor. He just thought we were close to self-driving in 2014.

Another factor is Andrej Karpathy, who was the primary architect for the vision-only approach. Musk wanted fewer parts, and Karpathy believed he could deliver that. Karpathy is still an advocate of vision-only.

Right, for the reasons that I just mentioned

> They also published the report here: https://pub-fee113bb711e441db5c353d2d31abbb3.r2.dev/VTT_CR_0...

That's the same link. Is there a way to attest that this is an official VTT report?

The person whose signature is on it, is listed on VTTs website with the same title.

Considering the amount of publicity this thing gets, VTT or the person will publicly refute it pretty soon if it's a fake.

Also, VTT had publicly confirmed earlier that they had conducted tests for Donut Lab.

I'm confident the document and tests are real, but other shenanigans are still possible (and likely IMO).

Sorry, I was trying to compose a post but I was a minute late so copy-pasted the content as a comment and forgot to adopt. Removing the redundant link. Thanks.

BTW, the people who conducted the test appear to be on LinkedIn. I guess its pretty easy to confirm if the test on the company site is authentic.

The PDF is digitally signed with a cert from the Finnish „Digital and Population Data Services Agency“

I think it is safe to assume, given the buzz all around the donut battery, that VTT would immediately release a statement if this report was fake.

edit: https://www.vttresearch.com/en/news-and-ideas/donut-lab-comm...

Unfortunately that's not the same as VTT.

VTT would be more like "National Institute of Scientific Research"

Well there's this press release they would publish a report: https://www.vttresearch.com/en/news-and-ideas/donut-lab-comm... with as author the same name on the digital signature "Petri Söderena" for Organisation "Teknologian tutkimuskeskus VTT" and the chain is attested by "DVV Organisational Certificates - G4E" which is on the EU/EEA trusted list: https://eidas.ec.europa.eu/efda/trust-services/browse/eidas/... (by name and key signature). Looks like a legit VTT document to me.

It says its sign by this guy: https://www.vttresearch.com/en/news-and-ideas/petri-soderena...

He has an e-mail address and a phone number, I doubt that if the report is falsified it won't come out.

Elon Musk sells self driving cars since 10 years that don't self drive but his cars are actually decent cars and his rockets are revolutionary. Also, who isn't selling magical AGI since the last 4 years?

I think i will judge the battery and the magical AGI separately. The guy also sells magical motors that appear to be real with people riding motorbikes with those motors.

> Oxfmt got it, and then won.

This doesn't align with the data.

  - 650k npm downloads for oxfmt this week [0]
  - 4.3m npm downloads for biome this week [1]

  - 19.1k gh stars for oxfmt [2]
  - 23.7k gh stars for biome [3]
  - gap widening in favor of biome [4]

  [0] https://www.npmjs.com/package/oxfmt
  [1] https://www.npmjs.com/package/@biomejs/biome
  [2] https://github.com/oxc-project/oxc
  [3] https://github.com/biomejs/biome
  [4] https://www.star-history.com/#oxc-project/oxc&biomejs/biome&type=date&legend=top-left

There's an active proposal for rendering html in canvas to control styling. Text rendering and accessibility are main focuses for the proposal.

https://github.com/WICG/html-in-canvas