Which non US government are you talking about. China is investing heavily and making at least some progress. EU couldn't get "legacy rocketry programs" working, so the reason is incompetence not stubbornness. India is trying a bit, but they don't have capacity to spend. Russia is also trying.
DeepSWE was released on May and Deepseek v4 flash was released on April. By definition it couldn't be benchmaxxed on this particular benchmark. My guess is the mutation is not like real world and it is tripping small models, not benchmaxed model.
It's not clear which v4 flash they used. There are 3 different releases for that one.
Edit: I think that's the answer. This page lists the score as 53.3% which is close to the 54.4% reported here https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731 for the 0731 update. The original from April was called preview, doesn't list the score, but in the comparison with the update it scores 7.3%.
The actual vulnerability got very little discussion. It's clever. Because Linux networking people are obsessed with checksum performance, virtio-net plumbs checksum offload metadata through the host/guest driver interface. Guests can thus deliver the host packets along with instructions on what to checksum. The vulnerability is a missing sanity check that allows checksum instructions to instruct the host to overwrite the header length field in the IPv4 header, rather than the checksum field.
Crucially, this happens in the middle of packet processing, so the kernel code is forced to work with two different ideas of how long the IPv4 header is. All hell predictably breaks loose.
Additional color: per the commit, the bug was found by Paulos Yibelo, who runs PwnAI (the author of this piece), and worked on security at AWS for 4 years prior.
I don't mind the company that makes me wait long, and in many cases prefer it. Just I reject any take home if takes more than few hours without communication, and I reject more than 1 round of interviews done by someone junior than me.
I like giving interviews. It gives me insight into their thinking and the company.
Most companies want the new joinees to be satisfied, and if the reason is "we don't have budget" that might be acceptable. But if the reason is "we are lowballing you specifically because we can" that would just create spite in candidates mind and the company has to pay the price for the process if a candidate dropped off or joined temporarily while looking elsewhere.
No question needs answering. If you want to run claude code inside sandboxed environment, it's not that hard and there are lot of options including official[0].
From the doc, this suggestion is hardly satisfactory:
The sandbox covers shell commands only. Claude’s file tools, MCP servers, and hooks run outside it.
Also this is in my view the main job of a scaffold like claude code - to restrict what the agent can see and make sure it doesn't use info it should not. It's not something a user can easily or should have to attempt to control themselves.
File tools, by default asks permission to edit outside directory. There is no MCP tools or hooks by default and if you install it you need explicit permission to trust it.
It does change the point! If it happened roughly twice a year, it's not as noteworthy when it happens than if tends to happen every other year, but four times.
It's more noteworthy when it happens the more clustered it tends to be. If it were typical that we'd go ~7 years with none then have ~16 in a single year, it would be huge news when it happens.
Corrections (down 10% or more): about one per year on average. The Nasdaq is more volatile than the S&P 500, which has a correction closer to every 1.5 to 2 years.
Firstly you are assuming $0.048$/hr for GPU, which is highly misleading and you can easily sell it for double(vast.ai is $0.12) as you are forgetting lot of cost like CPU, cooling etc.
Also even if you can somehow fit a few 100 billion parameter model(like gpt 6 luna or deepseek flash) in 16GB, and magically assume you can run at 200 tokens/sec according to your calculation, the blended cost of it would be around $100, not $1639.
reply