> Gabriel said Mossad had learned that Hezbollah was buying pagers from Gold Apollo, a company in Taiwan.
> "When they are buying from us, they have zero clue that they are buying from the Mossad. We make like the 'Truman Show,' everything is controlled by us behind the scene," Gabriel said. "In their experience, everything is normal. Everything was 100% kosher."
> To further the plot, Mossad hired the Gold Apollo saleswoman Hezbollah was used to working with, who was unaware she was working with Mossad. According to Gabriel, she offered Hezbollah the first batch of pagers as an upgrade, free of charge. By September 2024, Hezbollah had about 5,000 pagers in their pockets.
> Going analog has been a signature move for terror groups ever since the September 11th attacks as a way to successfully mask communications from Western militaries and government defense agencies.
> A source cited by the Wall Street Journal said many of the affected devices were from a new shipment delivered to Hezbollah militants in recent days.
> Apparently, the encrypted pagers currently in use by Hezbollah were brand new models and bought in bulk for the members just a few months ago, several sources told Reuters.
I'm sure you can find more if you look; there's a lot of articles about it.
Creating small, specialized models for specific tasks. Being able to leverage the up front training/data as a generalized base allows you to quickly create a small local model that can generate outputs for that task that can come close to or match the same you would see in a large/hosted model.
The market hype is real. Check-signers at businesses expect LLMs to have the ability the AI CEOs talk about in their interviews and conferences but don’t exist (and are no where near existing).
My understanding is it means half of what a subscriber pays is spent just on the compute required as you chat with the models. Which leaves the other have to be divided up among salaries, marketing, R&D, etc.
That’s all well and good, but how can we make money off of this post?
Well I’ve been working on an agentic app with a fun UI (https://aimakememoney.com) to turn complaints into startups ready for their first round of seed funding.
That's amazing. I've been working on a tool that takes seed money and turns it directly into apps, and thats been generating a ton of complaints from users. There's an unbelievable synergy between our two products, and I believe that if we combine forces we might be able to corner the user complaint market. Imagine the possibilities. Imagine the moat.
In other news, I would like to offer the domain aimakememoney.com to the person with the funniest thing to put on it. I'll consider submissions for the next 24 hours and choose a winner.
Current data scientist here, working for a cloud consulting firm. Two things stand out from my experience: (1) my company isn't hiring, while the DS team is doing fine revenue wise, the rest of the company is doing poorly; so uncommunicated moratorium on hiring; (2) I interviewed at an AI company that I'm currently subcontracting under - and they like me - and I didn't get past the first round because their requirements are so high right now (aka I did mediocrely on one interview and that was enough to tank me).
All this is to say, GenAI is booming but there's competing factors going on for businesses to hire.
Also a different take, look for contract jobs. As with (1) above, my company isn't hiring FT but they're open to contractors.
In my opinion, Nature is part of the problem in academic research. In the US, taxpayers must pay to access the information that they paid to fund! And then the researchers (other taxpayers) have to pay to publish and pay to access the content.
This manner of doing business is frustrating to me. Personally, Springer executive board, I hope this hits your bottom line.
Classes and workshops, something with the same people that occurs over several weeks. But it’s important that the content is something you’re personally interested in.
After reading the abstract, this title seems to me like flame/click-bait.
This test has been done in mice only (since no other animal was mentioned in the abstract).
If one follows the scientific method, this means whatever conclusion the authors make can only be extended to mice. Extrapolation to “mammals” is not present in the abstract implicitly or explicitly.
> Two experiments with different challenge-to-contact ratios were conducted to assess transmission dynamics and mutation development. In experiment 1, a 4:1 challenge-to-contact ratio resulted in 100% transmission among direct-contact mice, with all mice succumbing to the infection. In experiment 2, a 1:1 ratio yielded 50% transmission, with all challenged mice also succumbing.
It’s not clear from the abstract that any human-made changes were made to the H5N1 either.
That’s a tough take but I think fair. There’s also a common thought in industry that nothing gets done in academia (e.g., credentials are often weighed less from a university than from employment). Or at least anecdotally for me.
reply