jp0001

American models == Cars with no brakes. Chinese models == Cars with brakes that can't drive down tiananmen square.

I really don't see how the SaaS models will be profitable if this continues, all the money will be in providing consumers with hardware that can run these locally - with the ability to modify the weights and do your own ablation.

This is a sad. Makes me want to move my bank accounts.

If they paid for the tokens, then is it really stealing or just learning?

I uploaded a sample found here (https://github.com/alexct142010-cell/McBackuper ) to Genus Codes (need an account): https://genuscodes.com/results/7ad4b911d05a12f91ab27ba3baa35... Seems to be related to the disco trojan family, by way of normalized function matching at 50% to malicious file https://genuscodes.com/results/eddbc29db4677e00c1a901aadbadb... and a normalized 50% match to https://genuscodes.com/results/fdb6cff68a2a8c08779d64a7cf61d...

Virustotal link: https://www.virustotal.com/gui/file/fdb6cff68a2a8c08779d64a7...

Formal methods is like a plan. Everyone has one until they are punched in the face (real world requirements and trust boundaries).

Gotta have a moat.

Agree. If this was this was just "look at what Fable could do quickly" that's fine. But saying it's free for now is funny. Anyone else could just make games they want to play locally for themselves and their friends now. And there are better ones online.

It's time to make truly open source frontier models that people can run at home. Code is free speech. We've been through this with encryption algorithms in the past.

I asked it to write security tests for an app and I was downgraded to Opus 4.8. I'm approved for their cyber program!

You should throw GPT into the mix to UX/UI and call it the three stooges.

My guess in the North East in the winter there will be similar stories.

LLMs are going to produce amazing Rube Goldberg style vulnerabilities for years to come. It's already starting, this instance isn't the case, but it's happening.

That website was not for me.

Count 1 and 2 makes sense. A good lawyer can get 3-5 thrown out in a plea deal.

I'm still having problems trusting my compiler.

Max x20 user here. As long as Opus 4.6 is available and they fix Opus 4.7, I'll stay with Anthropic. Tho, I'd imagine in 5 years we'll have Opus 4.6 equivalent performance available in an at home consumer model.

It's easier to produce vulnerable code than it is to use the same Model to make sure there are no vulnerabilities.

WTF. `Opus 4.7 is the first such model: its cyber capabilities are not as advanced as those of Mythos Preview (indeed, during its training we experimented with efforts to differentially reduce these capabilities). We are releasing Opus 4.7 with safeguards that automatically detect and block requests that indicate prohibited or high-risk cybersecurity uses. `

Seriously? You're degrading Opus 4.7 Cybersecurity performance on purpose. Absolute shit.

I'm starting to think that Opus and Mythos are the same model (or collection of models) whereas Mythos has better backend workflows than Opus 4.6. I have not used Mythos, but at work I have a 5 figure monthly token budget to find vulnerabilities in closed-source code. I'm interested in mythos and will use it when it's available, but for now I'm trying to reverse engineer how I can get the same output with Opus 4.6 and the answer to me is more tokens.

I took three weeks off from tech, read books from last century, and travelled Europe. Coming back, reading LLM generated content and code feels like nails on a chalkboard. Taste, it does not have taste.

Hold on. Cell towers still know where the device is. If a group of people in an area have stable ismi’s and one person’s ismi is rotating daily, it doesn’t take a genius to figure out who’s now using cape. Using it for travel makes sense, but again being a device that doesn’t a have an owner is, as the kids say, sus.

Mobile device updates are the worst for aging parents. These devices are getting more complex to use not easier, you shouldn't have to upend your life once a year because UX design choices forces you to miss what you think is important, how to find it, or disable/enable features you don't want or used to have.

Eh, looks at my aws bedrock bill, I think they are doing alright.

Every time I read something like this, I remember that there is truly no correct way to say something - all that matters is that your intended audience understands it, eventually.

You don’t think LLMs should get paid for their labor? /s

So sick of this. LLMs are just a new paradigm on search.

Old Google: I don’t know the answer, but here are some links that might help you - with ads.

LLMs: I confidently know the answer, but I could be wrong and frequently am. No ads yet, but they are coming.

I want to know how much they made in bounties versus how much they spent on compute.

The thing about bug bounties, the only way to win is to not play the game.

Correct. Dynamic runtime interactions will always be a hard problem as it’s hard to see in static code even for humans.

We followed a very similar approach at work, created a test harness and tested all the models available in AWS bedrock and the OpenAI. We created our own code challenges not available on the Internet for training with vulnerable and non-vulnerable inline snippets and more contextual multi-file bugs. We also used 100 tests per challenge - I wanted to do 1000 test per challenge but realized that these models are not even close to 2 Sigma in accuracy! Overall we found very similar results. But, we were also able to increase accuracy using additional methods - which comes as additional costs. The issue I see overall is that we found is when dealing with large codebases you'll need to put blinders on the LLMs to shorten context windows so that hallucinated results are less likely to happen. The worst thing would be to follow red herrings - perhaps in 5 years we'll have models used for more engineering specific tasks that can be rated with Six Sigma accuracy if posed with the same questions and problems sets.