Lightnews — Scholar-powered news

Reposted by Stefano Zanero

Ed Zitron @edzitron.com · 1d

Jesus christ. AI coding platform Augment code had to jack up prices because 22.5% of their users were spending 20x of what they paid, and even after raising the prices, they're still running at a loss.
reddit.com/r/AugmentCod...

A handful of users abused the system so all are getting punished.

This isn't about a few high-usage users. The reality is that approximately 22.5% of our users are consuming 20x what they're currently paying us. This isn't sustainable for us to continue delivering the quality service you expect. We have built some very powerful tools and we don’t want to impose artificial limits on what’s possible, but we do need to be able to charge in proportion to the use customers are getting from our platform. Developers are always going to push their tools to their limits, and we encourage that — and we need to be able to charge for it appropriately, too.

You only care about professional developers.

Our core focus is on building the best AI coding agent for professional software engineers and their teams. If people outside of that group are finding value with Augment, they are very welcome to use the product, but we’re not prioritizing features or solutions that non-developers might need, and frankly, there are plenty of vibe coding/low code/no code solutions available on the market that will better serve those customers.

You are just squeezing money out of us at 20x margin.

20x margin sounds great, but isn’t the reality for AI tools: the vast majority are running at a loss, including us, while we work to build sustainable, long-term businesses.

57 300 1.6K

Reposted by Stefano Zanero

Techmeme @techmeme.com · 5d

Apple announces a "major evolution" of its Security Bounty program, doubling the top award to $2M for exploit chains that could be abused for spyware attacks (Lily Hay Newman/Wired)

Main Link | Techmeme Permalink

3 9

Reposted by Stefano Zanero

Jeremy Blackburn @mrjimmyblack.com · 3d

It’s amusing to see people that clearly have no technical understanding of how LLMs work pontificate on how they work. They’ve clearly picked up some of the appropriate language/concepts, but it’s uncanny valley level stuff where their use is off in subtle ways that expose that they have no clue.

2 1 8

Stefano Zanero @raistolo.bsky.social · 3d

Wat

Mark Lemley @marklemley.bsky.social · 4d

The crash is going to be brutal when it comes.

Robinhood says "it might offer "tokenized" equity in OpenAI. By "tokenized," Business Insider reports, Robinhood means "blockchain-enabled representations of securities like stocks." In reality, they have no connection to OpenAI equity whatsoever."

3 1 8

Reposted by Stefano Zanero

Meredith Whittaker @meredithmeredith.bsky.social · 6d

We're so relieved to see Germany reaffirm its opposition to the dangerous Chat Control proposal--the one that would mandate mass scanning of communications. Germany's long been a solid champion of privacy, and the news that it was considering backing mass surveillance was alarming. 1/

9 220 1K

Reposted by Stefano Zanero

NightMar on Hicks Street, Part 4: The Rabbit’s Revenge @histoftech.bsky.social · 6d

“Concerns over an AI bubble bursting have grown lately, with analysts recently finding that it’s 17 times the size of the dotcom-era bubble and four times bigger than the 2008 financial crisis.”

Hang onto your butts. This “correction” is gonna hurt.
futurism.com/artificial-i...

Bank of England Warns of Impending AI Disaster

The Bank of England has sounded the alarm, warning of an intensifying risk of a "sudden correction" due to an AI spending frenzy.

futurism.com

42 460 1K

Reposted by Stefano Zanero

Stephen Hampshire @shampshire.bsky.social · 5d

I have endless circular arguments about this.

Them: “It’s written a summary.”

Me: “No, it’s written something tuned to look like a summary.”

Them: “But it looks like a summary.”

Me: <sigh>

We’re not used to computers lying to us.

7 18 85

Reposted by Stefano Zanero

Matthew Green @matthewdgreen.bsky.social · 5d

Germany has agreed to stop ChatControl for now, due to huge amounts of public pressure. Good job! The bad news is that it could come back as soon as December, and the German government has interpreted the feedback as a need to “moderate” the proposal.

1 47 150

Stefano Zanero @raistolo.bsky.social · 5d

Nah.

1 2

Stefano Zanero @raistolo.bsky.social · 5d

Possibly. On the other hand, this makes AI doomerism incredibly misguided.

1 1

Stefano Zanero @raistolo.bsky.social · 5d

“GPT-5 can do things no other A.I. can do. It can hack into a web server. It can design novel forms of life. It can even build its own A.I. (albeit a much simpler one) from scratch.”

None of this is true.

It cannot “do” any of those things. At most it can help experts do them. And even there…

1 1 1

Stefano Zanero @raistolo.bsky.social · 5d

L’aviazione è un ottimo modo per essere poveri ma felici

2 6

Stefano Zanero @raistolo.bsky.social · 5d

Basta decidere il budget, poi raddoppiarlo e guardare cosa si può comprare 😂

1 7

Stefano Zanero @raistolo.bsky.social · 5d

“Use this technology that has been proven to make your work slower overall to be 5 times faster in building the other technology that no one actually wanted and everyone just pretended never happened”

404 Media @404media.co · 5d

NEW: an internal Meta message we obtained tells employees on its metaverse team that they should be using AI to “go 5x faster”. That's 5 times faster, not 5 percent more www.404media.co/meta-tells-w...

Meta Tells Workers Building Metaverse to Use AI to ‘Go 5x Faster’

Meta says that its coders should be working five times faster and that it expects "a 5x leap in productivity."

www.404media.co

2 3 16

Reposted by Stefano Zanero

Armando Bellante @ikiga1.bsky.social · 6d

I’m happy to announce a new #preprint! 🧑‍💻📝🎉

Quantum states often show up with hidden structure. What if a state is built from just a few elements of a larger, #non-orthogonal, #overcomplete dictionary? Can we exploit that sparsity to beat standard #tomography costs?

🧵⬇️ /n

1 1 3

Stefano Zanero @raistolo.bsky.social · 14d

Ma a parte tutto… “nuovi inoculi”? Iniezioni pareva brutto? Bisogna trovare termini fantasiosi per farle sembrare cose più strane?

1

Stefano Zanero @raistolo.bsky.social · 15d

That’s true only in a few airports around the world (it’s called preclearance, and it has happened for at least 15 years)

1

Stefano Zanero @raistolo.bsky.social · 15d

No. It’s most definitely not. This is ridiculous. Absolutely ridiculous.

The Guardian @theguardian.com · 15d

It’s time to prepare for AI personhood | Jacy Reese Anthis

Technological advances will bring social upheaval. How will we treat digital minds, and how will they treat us?

www.theguardian.com

1 4 18

Reposted by Stefano Zanero

Luke Plunkett @lukeplunkett.com · May 1

jesus christ

23 570 1.9K

Stefano Zanero @raistolo.bsky.social · 15d

Memories.

Jason Scott @textfiles.com · 15d

Nobody asked for it. We did it anyway.

If you ever used Winamp during the later years, it had an awesome, awesome visualizer called Milkdrop. The Webamp people implemented a workalike called Butterchurn.

Internet Archive now supports Webamp and Butterchurn. Just click on the llama!

1 6

Stefano Zanero @raistolo.bsky.social · 15d

That’s a defcon talk, not a black hat talk.

2

Stefano Zanero @raistolo.bsky.social · 15d

Drop by!

Stefano Zanero @raistolo.bsky.social · 16d

Non dirlo mai più. Mai. Più.

1

Reposted by Stefano Zanero

Michael A. Specter 👻 @mikespecter.com · 16d

Today, my research group @ Georgia Tech released a paper on vulnerabilities in Tile --- the second largest device finding network after Apple's AirTags.

You can read about it in Wired, reporting by @kimzetter.bsky.social!
www.wired.com/story/tile-t...

1 34 61

Stefano Zanero @raistolo.bsky.social · 18d

a knight with a helmet and a sword says " this is the way "

ALT: a knight with a helmet and a sword says " this is the way "

media.tenor.com