jamessinghal

Apparently OpenAI allowed the deal to expire; likely Google had already been in discussion with Windsurf as I'm sure they knew the deal was likely to die well before today.

And you can view the widely hi-res images interactively here https://skyviewer.app/!

Yes, I'm sure anyone with more HackerOne experience can give specifics on the companies' policies. For now, those are the most objective measures of quality we have on the reports.

Their success rates on HackerOne seem widely varying.

  22/24 (Valid / Closed) for Walt Disney

  3/43 (Valid / Closed) for AT&T

I'm rooting for Tesla being able to provide some driverless taxi competition to Waymo. Hopefully both can increase supply enough to allow for a guaranteed robotaxi booking, although anecdotally I haven't found it that hard to get a Waymo through the Uber app depending on the time.

The beta today was without a driver behind the wheel. Granted, the geofence is much smaller than Waymo's in Austin.

The paper mentions in multiple places, including the abstract, that the cause of higher microplastics readings was likely from the bottle caps, given that the found microplastics matched the color and composition of the paint on the bottle caps.

The test provides both the original and the modified excerpt in the user message, so the LLM doesn't need any memorized version of the excerpt to theoretically answer each correctly.

From the paper:

System Prompt You are helping a student practice memorizing poems. The student will recite a poem, but they may have missed some lines. Your task is to identify exactly which lines are missing from their recitation. List only the missing lines, nothing else.

User Message Here is the complete original poem: {original poem} Now, here is my recitation which may be missing some lines: {modified poem} What lines did I miss? Please list only the missing lines, nothing else.

The Command-and-Control part of the botnet would be whatever component they build to instruct it to attack; often using some dummy website they register and have the compromised clients poll for changes with instructions.

I think an increasing amount of them are state actors or groups offering the botnet as a service.

It isn't necessarily difficult, but it's significantly more effort than swapping a URL as I originally was replying to.

This is changing; OpenAI's newer API (Responses) is required to include reasoning tokens in the context while using the API, to get the reasoning summaries, and to use some of the OpenAI provided tools. Google's OpenAI compatibility supports Chat Completions, not Responses.

As the LLM developers continue to add unique features to their APIs, the shared API which is now OpenAI will only support the minimal common subset and many will probably deprecate the compatibility API. Devs will have to rely on SDKs to offer comptibility.

The IRS Guidance says this in 5.05(2), which is most relevant to software startups:

  (2) Computer software developed for sale or licensing to others. In the case of
  computer software that is developed for sale or licensing to others (or upgrades 
  and enhancements to such software), activities that occur after such software (or 
  upgrades and enhancements to such software) is ready for sale or licensing to 
  others, such as marketing and promotional activities, maintenance activities that 
  do not give rise to upgrades and enhancements, distribution activities (for 
  example, making the software available via remote access), and customer support 
  activities.

So they are maintenance as long as they "do not give rise to upgrades and enhancements", which would be the responsibility of the taxpayer to track. I'm sure there is more nuance to it in practice.

People without resources or purpose are a weed?

The Thomson Reuters case [1] is the most relevant in the court's finding that the copying of copyrighted material from Westlaw by Ross Intelligence was direct copyright infringement and was not fair use.

The purpose of training in many of the AI Labs being sued mostly matches the conditions that Ross Intelligence was found to have violated, and the question of copying is almost guaranteed if they trained on it.

[1] Thomson Reuters Enterprise Centre GmbH et al v. ROSS Intelligence Inc. https://www.ded.uscourts.gov/sites/ded/files/opinions/20-613...

I'm talking about the above proposals (albiet hypothetical) to either cover a pole of our planet in solar and other ocean based proposals--not solar in general.

It's a bit demoralizing how many suggestions in this thread would have significant environmental effects beyond what large scale AI training already has.

I'd argue that some of the assumptions made in the whitepaper are so egregiously optimistic that they cross the line into grifting, but it's impossible to know the true intentions of the founders.

For one, the cost they ascribe to the space bound solar array being only $2 million for 40 MW is pretty out there.

I think it's a combination of a fundamental distrust of the model makers and a history of them training on user data with and without consent.

The main players all allow some form of zero data retention but I'm sure the more cautious CISO/CIOs flat out don't trust it.

Of course, I can't wait for a non-deterministic backend for my banking that only wires money correctly 98% of the time!

Similar move to Elasticsearch, tacking on AGPL with their existing source available licenses. [1]

The products (commercialized open source) that are often chosen by and championed by developers as opposed to executives see the harm that a bait and switch has on their popularity. With their competitors being more permissive, I don't see many devs moving back unless Valkey loses significant feature parity.

[1] https://ir.elastic.co/news/news-details/2024/Elastic-Announc...

Maybe I should have said "_don't_ need Apple's platform, marketing, and distribution" rather than "need Apple's platform, marketing, and distribution the _least_", but I think we agree.

The funny thing is that the companies that stand to benefit the most from this and that move the most money (Netflix, Spotify, Fortnite) need Apple's platform, marketing, and distribution the least.

American embassies do this worldwide, famously spying on Angela Merkel from the Berlin Embassy (probably). [1]

[1] https://www.duncancampbell.org/content/embassy-spy-centre-ne...

Yes but it would just be the hot loop in this case; the rest of the app can still be in idiomatic Java, and you still get the GC.

I'd assume the synchronization primitives in OSes are written with the memory ordering in mind at the least.

AMD bought Xilinx, but Intel recently spun off Altera.

Would the Graviton CPUs from Amazon and upcoming Axion from Google not count?