Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
lucamark
searching Neon…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
8 ms
·
1.
▲
by
lucamark
2mo ago
Exactly! And they should also release the new inference engine in this month. Anyway, I am curious to try Opus 5, considering that previous versions (e.g., 4.8) were disappointing
2.
▲
by
lucamark
2mo ago
But why GPT 5.6 Sol is so behind on the benchmarks? In real-world projects, it is the best frontier model to me in terms of accuracy, speed and consistency. It can just be compared to Fable 5, but I prefer GPT 5.6 Sol because of inference s
3.
▲
Agentic Misalignment in Summer 2026
(alignment.anthropic.com)
3 points
by
lucamark
2mo ago
|
0 comments
4.
▲
MDL – a text-based interface for visual control models (from an HN comment)
(modeloop.app)
2 points
by
lucamark
2mo ago
|
1 comments
5.
▲
by
lucamark
2mo ago
A month ago I posted Modeloop here, a visual environment for designing control systems ( https://news.ycombinator.com/item?id=48543968 ). In the comments,boscillator wrote: "I'm always saying Simulink needs a text-b
6.
▲
by
lucamark
3mo ago
For firmware RE this would be most useful if every guess is tied to evidence
7.
▲
by
lucamark
3mo ago
Using power as feedback is neat because it works when you can’t instrument the target but I wonder how they separate new behavior from boring analog noise, e.g. the clocks
8.
▲
by
lucamark
3mo ago
I’d like to see which parts of the assurance case are language/toolchain dependent versus plain old process, tests and traceability.
9.
▲
by
lucamark
3mo ago
Remember that such models are available to us thanks to NVIDIA GPUs
10.
▲
by
lucamark
3mo ago
Never had rejections in the short time Fable was available
11.
▲
by
lucamark
3mo ago
Agree. It is a mediocre model, expensive while not being a frontier
12.
▲
Claude Sonnet 5 – benchmark results
(artificialanalysis.ai)
41 points
by
lucamark
3mo ago
|
17 comments
13.
▲
by
lucamark
3mo ago
Wrong! Look at it better. It shows that Opus has superior performance but at higher cost.
14.
▲
by
lucamark
3mo ago
You're referring to the Agentic search, but if you look at the Agentic computer use the cost is basically halved. However, I am also confused about market positioning. Too expensive to perform daily tasks - open souce models are much c
15.
▲
The Key Benefits of Model-Based Design
(modeloop.app)
2 points
by
lucamark
3mo ago
|
0 comments
16.
▲
by
lucamark
3mo ago
I checked before posting but I didn't find it
17.
▲
How to Earn a Billion Dollars
(paulgraham.com)
13 points
by
lucamark
3mo ago
|
4 comments
18.
▲
Tesla FSD v14 Lite
(twitter.com)
2 points
by
lucamark
3mo ago
|
1 comments
19.
▲
by
lucamark
3mo ago
OpenRouter?
20.
▲
by
lucamark
3mo ago
> In other words, the Trump administration appears to be pressuring OpenAI to do what Anthropic is already voluntarily doing: keeping its most powerful AI models under wraps.
21.
▲
by
lucamark
3mo ago
while I generally agree that models with this level of cyber capability should not necessarily be released to everyone on day one, if a company or developer can meet compliance requirements, they should be eligible - and eligibility should
22.
▲
by
lucamark
3mo ago
Uh great! I haven't seen these posts, thanks for the links :)
23.
▲
The White House is asking OpenAI to slow roll the release of its new model
(techcrunch.com)
49 points
by
lucamark
3mo ago
|
14 comments
24.
▲
by
lucamark
3mo ago
Expections occur
25.
▲
by
lucamark
3mo ago
In many other european countries, the story doesn't change much - with some exceptions. The irony is that the EU is now discussing EU Inc. a law to let startups incorporate online within 48 hours, for at most €100, under a single optio
26.
▲
by
lucamark
3mo ago
Do you think that stuffs are different in Italy, Spain or France? Just to cite some others
27.
▲
by
lucamark
3mo ago
Europe will suffocate under the weight of its own bureaucracy. The sad thing is that it is not new to me, I've heard so many stories like this one. This is the kind of friction that makes founders incorporate elsewhere. Also, a founder
28.
▲
GPT-5 helped an immunologist solve a 3-year-old mystery
(openai.com)
2 points
by
lucamark
3mo ago
|
2 comments
29.
▲
by
lucamark
3mo ago
That's bad for ETF owners too. Funds will be legally forced to buy a tightly restricted supply that will pump the price even more
30.
▲
by
lucamark
3mo ago
Let's grab some popcorn. - Price/Sales about 128x (NVIDIA had a peak of a 40x at its peak) - Bought Twitter per 44 billions. Inflated its valuation to $250 billion just by integrating it into X.AI - EnterpriseValue/EBITDA abo
More ›