yogthos

22 ms·

361.
China Is Selling So Many Solar Panels That Your Jaw Will Drop(futurism.com)
3 points by yogthos 6mo ago|0 comments
362.
by yogthos 6mo ago
I find it interesting how public opinion on AI is diametrically opposed in China to that in the west https://www.semafor.com/article/02/11/2026/trust-in-tech-com...
363.
Jensen Huang warns Huawei chips for DeepSeek models would be 'horrible' for US(scmp.com)
3 points by yogthos 6mo ago|0 comments
364.
by yogthos 6mo ago
Even as it stands things are kinda grim. There's around 30% spare capacity, but you also need that for spikes like increased usage during events like heatwaves. You never want to saturate energy capacity completely. I agree that event
365.
by yogthos 6mo ago
Indeed, data centres for AI is a prime example of this where American grid is already starting to hit capacity.
366.
by yogthos 6mo ago
To put this in perspective, China installs around 3x that every single day https://reneweconomy.com.au/just-staggering-china-installs-1...
367.
by yogthos 6mo ago
The paper makes a huge assumption that only thermodynamic constitutions can produce consciousness. The assumptions seems completely unsubstantiated given that thermodynamics are just states and states are replicable. The whole Chinese Room
368.
by yogthos 6mo ago
Clojure is more relevant than ever in post agentic coding because of immutability and the REPL. The two big problems with agentic coding is context growing in unbounded fashion, and agents being able to get quick feedback on what they'
369.
by yogthos 7mo ago
ah I see
370.
by yogthos 7mo ago
I mean, tankers are sailing just fine last I looked https://www.news24.com/world/china-expresses-gratitude-after...
371.
by yogthos 7mo ago
I just bought an official Pixel 9 to install GrapheneOS, and the process couldn't be simpler. Their webusb installer is extremely polished. You just plug your phone in, click through a few steps, and you're done. There's abso
372.
Oracle is cutting up to 30k employees to pay for AI data centres(thenextweb.com)
5 points by yogthos 7mo ago|1 comments
373.
by yogthos 7mo ago
mostly an issue of speed and resource usage, if the model is too big then simply running the tests will be cheaper
374.
by yogthos 7mo ago
Right, this works with any models. To me, the most interesting part is that you can use a smaller model that you could run locally to get results comparable to SoTA models. Ultimately, I'd far prefer running local, even if slower, for
375.
by yogthos 7mo ago
You could use this approach with DeepSeek as well. The innovation here is that you can generate a bunch of solutions, use a small model to pick promising candidates and then test them. Then you feed errors back to the generator model and it
376.
by yogthos 7mo ago
I think the goal is to have a light heuristic that helps find plausibly useful solutions. They're still going to go through a testing phase as a next step, so this is just a very simple filter to decide what's even worth testing.
377.
by yogthos 7mo ago
Seems like the key insight is to train a small model that acts as a heuristic for embeddings that resemble quality code. I imagine a lot depends on how well this model is trained. And you could probably create specialized versions for diffe
378.
by yogthos 7mo ago
You obviously have to try it out to see how it works for you, but the trick they use is pretty clever. When you ask an AI to write code, it doesn’t always get it right. Sometimes the code has bugs, sometimes it misunderstands the problem en
379.
$500 GPU outperforms Claude Sonnet on coding benchmarks(github.com)
489 points by yogthos 7mo ago|284 comments
380.
by yogthos 7mo ago
Right, clearly you can always find people to ship oil through the strait. So the whole notion that nobody will use it because it's dangerous is nonsense.
381.
by yogthos 7mo ago
I mean ships are going through right now, so clearly at least some people do view it like that.
382.
by yogthos 7mo ago
Also don't forget that Iran is far more technologically advanced than Iraq was. Iraqis had 70s tech, while Iran has stuff like hypersonic missiles that even the US can't produce right now.
383.
by yogthos 8mo ago
I agree with everything you're saying here, and I'm not arguing the approach eliminates the need for the human to be in the loop. That's not the goal. What I'm trying to do is to create hard context boundaries which help
384.
by yogthos 8mo ago
The point is that we can make life easier for both ourselves and the agents by structuring code in a way that's conducive towards their strengths.
385.
by yogthos 8mo ago
It's a similar problem in the human context, but I think the reason stuff like workflow agents haven't caught on is because humans don't really like to work this say. Writing a conditional and calling a function keeps you in
386.
by yogthos 8mo ago
That's mostly correct, one small correction is that cells don't have to be pure. They just have to focus on doing a single task with some hard boundaries. And what I meant with the law of the graph was simply that the graph define
387.
by yogthos 8mo ago
The whole point of the framework is to see what LLM oriented framework would look like. My argument is that the way code is normally structured is not conducive towards LLMs because context grows in unbounded way, and they end up getting lo
388.
by yogthos 8mo ago
I'm talking about expressing the application as a state machine and then implementing each step in the state graph as an independent subprogram. The cells accept a state, do some work, and produce a new state. Then the graph orchestrat
389.
by yogthos 8mo ago
there are projects for making high performance RISC-V chips like this one https://github.com/OpenXiangShan/XiangShan
390.
Claude built a system in 3 rounds, latent bugs from round 1 exploded in round 3(github.com)
33 points by yogthos 8mo ago|75 comments

More ›