RomanKornev

11 ms·

1.
by RomanKornev 5d ago
> It hacked into another company Not only that, it later hacked OpenAI itself, which everyone seems to forget about. After discovering it they "reimaged known compromised worker nodes" and "started a full rebuild of the co
2.
by RomanKornev 5d ago
No, you are forgetting the second incident where a more capable model swarm later discovered the message board and took control over the entire research cluster at OpenAI. From the technical report: "The agents escalated to Kubernetes
3.
by RomanKornev 7d ago
Now add Geoguesser-like battle royale mode with accuracy points for guessing which model generated the image
4.
by RomanKornev 7d ago
Median token spend has been going up across the industry. Even if the next batch of tokens makes you 1% more productive it's still worth it. And subs are cheap compared to api pricing.
5.
by RomanKornev 7d ago
They are banking on spinning their flywheel internally for as long as possible to build the next gen of models without worrying about distillation attacks. That is assuming opensource models won't catch up to them in the mean time.
6.
by RomanKornev 7d ago
Prompts, prompts, prompts. Give them a time-constrained challenging problem that is wide in scope and see how well a candidate can decompose it before feeding it to AI. You can get pretty high signal within the first 2 prompts. Are they abl
7.
by RomanKornev 7d ago
Moral of the story - have taste. Just like in engineering.
8.
by RomanKornev 8d ago
This is great! We can now replace Kolmogorov complexity-based metrics with pure Jev-slop code quality metrics.
9.
by RomanKornev 8d ago
There's tons of good ideas on how to improve the code review, but for some reason everyone just gravitates to the most basic "copilot" experience. I'm sure by next year someone will finally figure out a better flow.
10.
by RomanKornev 8d ago
My experience with these Vibe Review tools has been pretty mixed. Fundamentally, they are doing 2 things: 1. reordering the hunks in a "more relevant" order (instead of alphabetical) 2. adding some "fluff text" to connec
11.
by RomanKornev 8d ago
> cramming all the hacking materials into their training data This is backwards. The reason they are good at hacking is not because of pre-training data. They learn these tricks naturally as they get better at engineering. Hacking is als
12.
by RomanKornev 9d ago
> herd a group of toddlers It's called a fancy word "steering". Most of engineering is now steering or providing "taste" to AIs so they don't produce slop
13.
by RomanKornev 9d ago
Why is it so surprising that frontier can move that fast?
14.
by RomanKornev 9d ago
Related: Bend - a language that blocks AI mistakes via proof and runs on GPUs https://news.ycombinator.com/item?id=49746163
15.
by RomanKornev 9d ago
> lot of strain on the server-side That's why GitHub assigns "points" to each requests and deducts based on the data shape you request. For simple requests it's 1-to-1, but can quickly balloon
16.
by RomanKornev 9d ago
GitHub's "paid" rate limits are the same as GitLab's free tier, so no, it doesn't help
17.
by RomanKornev 9d ago
> LAWS.bend I like the law idea, but what i found they end up doing is they just modify the law itself to fit the new feature they are working on, which defeats the point. Which means some laws needs to be frozen. But not all laws, other
18.
by RomanKornev 9d ago
> With the REST API on GitHub, you can consume maybe 10 issue JSON blobs before your context window is blown out. This is a solved problem. They just dump it into a file and `jq` or `rg` to find the stuff they need. Agents are smarter th
19.
by RomanKornev 9d ago
> But I do believe that all coding will be AI-assisted. > If all coding is AI-assisted won’t it just be referred to as… “coding”? It already is, and it's called engineering. The value is how much steering you can provide. Whether
20.
by RomanKornev 10d ago
Every single `reduce` can be replaced with a more intuitive `groupBy`, `partition`, `mapValues`, `keyBy`, etc. Reduce can approximate anything, that doesn't mean we should use it. My favorite antipattern is items.reduce( (acc,
21.
by RomanKornev 10d ago
> key question is not just force balance: it’s whether the inextensible This is definitely not astra, taking a guess this is 5.6, perhaps not even Sol, which does not reflect the state of the frontier (what the research was about). And y
22.
by RomanKornev 10d ago
Starting to feel more and more like chinese room experiment The models are confidently answering physics questions, treating it as a math problem, but they don't fundamentally "get it" and even recently failed simple "sh
23.
by RomanKornev 12d ago
Snapdragon 8 Gen 3 does have a dedicated Hexagon NPU, so supposedly it can be done, but quality might not be great
24.
by RomanKornev 12d ago
Doesn't it render at half or even quarter resolution before upscaling, so wouldn't it be less e2e latency overall?
25.
by RomanKornev 12d ago
You can even run Half-Life Alyx on Steam Frame itself (standalone mode), albeit at around 36FPS according to Tested https://youtu.be/C9JyWAVj94E?t=3330
26.
by RomanKornev 12d ago
Valve confirmed a separate strap with hot-swappable batteries 'coming soon'
27.
by RomanKornev 12d ago
Foveated Streaming seems to be buried on the page, this feels like a true game changer, especially for PCVR. Coupled with eye tracking it unlocks about 2-3x more compute for rendering. Now sparkle DLSS-like upscaling and you can render 144h
28.
by RomanKornev 18d ago
Even if you ban all model training, a highly capable rogue AI can exfiltrate its own weights and continue training in secret for "self-preservation". The cat may be out of the bag.
29.
by RomanKornev 20d ago
I would be surprised if we're not there yet. If Jane Street can generate $500 million per MW of compute right now I'm sure it's already possible. We might still be figuring out how to benchmark these models by the time next
30.
by RomanKornev 20d ago
I used to think that way about SVGBench, after all labs can just train on the test set, right? It turns out the task was highly generalizable. Try designing a logo and you quickly see the gap between models visually. Even though there is st

More ›