If CoreWeave is likely toast ... news.ycombinator.com/item?id=4592...

you impression that, from time to time, you have terrible ideas may come from the fact that asking people about those ideas made apparent they were terrible immediately. Its not you, its just that ensembling works. Next time you will be the sane one.

Sounds interesting! luma.com/agentic-fron...

Reading people talking about Coding Agents and boy things are improving really fast. People digging out awesome manuals to feed to an agent to make it better ... news.ycombinator.com/item?id=4591...

Unlike ICLR reviews, the sunrise never disappoints

Happy There Is A Circulatory System Walking Through The Kitchen Day to all who celebrate.

Three dumb ideas that recur constantly in discussions among nerds about intelligence: (1) IQ tests are illegal in hiring. (2) There exist national average IQ numbers. (3) Intelligence research is suppressed.

For anyone that has worked in DL long enough to go from DIY to theano/chainer/tensorflow/mxnet/dynet/pytorch/Jax, one of these "winning" over all other felt unlikely ... and it kinda happened

For machines we call it "Reward Hacking" for humans we call it "Goodhart's Law". It makes us look less stupid (only look).

It's not going away folks ... ok, there are some scenarios, but on those something else went really wrong and we have far serious problems to think about

Another shot of the xpeng robot. really surreal stuff.

Claude Code's biggest flaw is that the cursor on the cl does not blink! When jumping around in tmux I often end up inputting things into CC that were meant for some other pane/window. Generally exits, which Claude understands in many languages ...

Guessing which keywords can I pronounce properly in every language to get Google's automatic language detection ASR to pick the damn right language. For Portuguese, "Oi" seems to do the trick.

Ok, time to confess. I do not gamble, so the second time I re-joined Twitter I did it with the express intent of using it as a prediction market so that my tendency to make claims bore some risk/reward. This one aged well.

Is there something fundamentally wrong with this reasoning?

Here is an argument why the place in the supply chain with the biggest value capture potential for the upcoming AI industrial revolution is neither in LLM training, nor infrastructure. Yeah, space is crowded, requires gigantic investments that deprecate fast , but also 👇

apparently Logan is on bsky now?

Observational tokenonomy

This is a collab with DeepMind & Terence Tao btw

CYBERSPACE 2 for(float i,z,d;z+i++<7e1;o+=vec4(z,1,9,1)/d) {vec3 p=abs(z*normalize(FC.rgb*2.-r.xyy));p.z+=t*5.;p+=sin(p+p);for(d=0.;d++<9.;p+=.4*cos(round(.2*d*p)+.2*t).zxy);z+=d=.1*sqrt(length(p.xyy*p.yxy));} o=tanh(o/7e3);

This is all the same guy??

Chris Lattner is one of the most influential engineers of the past two decades. He created LLVM, Swift, contributed to TensorFlow, and created the Mojo language. What was the story about creating Swift - and why did he face resistance inside Apple when wanting to replace Objective C? (cont'd)

Too big too frail

So something out there seems to have blown up at scale so massive we don't even have a good explanation for? Like ... star exploding while devoured by black hole falls short, ok ...

Every image this account posts would have sold me a paperback book in the 1980s. I would have been unable to resist these covers.

It's going to be a difficult few days for racists. No sooner do we discover the attacker was a British man called Anthony Williams, we then discover the rail worker hero who almost gave his life to save others is called Samir Zitouni. We can't wait to hear the hot takes on whether he is British...

Sort of starting to believe that we really do need academic metrics that punish publishing too much

I honestly cannot understand why modern ASR engines do not exploit the "burstiness" of language. Okay you failed the first time I pronounced poorly that foreign name or weird term ... but now you know I'm most likely going to use it again!

i.e. it's kinda horrible, but one should look at unemployment not only in terms of "number of people" but also in terms of "number of dollars" that are not in consumers hands to keep the service sector and other parts of the consumption economy alive.

There is a good anti-AI case to be made but it's mostly none of these. It does work and is much better than anything we had. Yes there is hype, but now that it is productized people can test by themselves (it's free!) and the standard startup hype is endemic to whatever the hot topic is 👇

Wow, a comment on HN I actually agree with!

Just checked on ClearSky and I'm not on a single blocklist and now I'm worried that I'm not influential enough.

Lore on why BlueSky keeps the low word count.

Lots of great ideas are simply figuring which features are people willing to give away in exchange for other features. You can't have "luxury low cost airlines", but this won't stop people from asking for them.

News from the LLM companies vs Neuro-Software companies front. LLM ones own the upstream and can optimize training to their products. NS companies have domain knowledge but more limitations on applying it. However ... they can optimize for runtime infrastructure, which they own!. Maybe this is key?

When you're late for your $500 monthly payment for the third time

A brief explainer about the 1x Neo since people are confused. The Neo has two modes: autonomous and teleop. The autonomous mode works for some tasks. In the teleop mode, for tasks it doesn't yet know, you can schedule a remote operator to operate it. They are remarkably upfront about this!

TIRAL you fix a bug in your experiments and results get worse. What's going on? Does the second law of thermodynamics not apply to you? Nopes, your hyperparam tuning had compensated for that error already ... so when you retune results better be clearly above the original results in your metrics

IBM is cooking, apparently Granite-4-Nano is a 1B model that beats Qwen3 1.5B huggingface.co/blog/ibm-gra...

Wait is this true? LOL x.com/m2saxon/stat...

www.arxiv.org/abs/2510.18212 ... meanwhile github.com/microsoft/ca...

You are low surprisal. Worst kind of insult.

If you’re going to post a paper on twitter, why not do it a few days after the bluesky post? No harm to your career but makes clear it’s a slower information source

So. much. slop. in YouTube

Battery slot as big as its chest

RE: these posts about RL screwing Claude's coding style. For code Gen standard RLVR would use cheap "it runs" unit tests plus expensive human created/curated "solves problem" ones. First can be reward-hacked by silencing errors, but not the second. Maybe it's a matter of data balance/money/time? 👇

Half of the images on this article are straight out of a Clark Ashton Smith short story

One of our students vibed an extremely useful map visualizer in a few hours. We would just never have done it before because it would have been a week of work. We are in a new world.

Oh jeez Slavoj Žižek has aged and his "answers" have even higher branching frequency. He is never going to return to the interesting point in bifurcation number 4, right?

TIL if you are discussing math with Claude Code (you should) you can ask it to use "pdflatex" and "open" to render a canvas that you can put on a split screen and ask it to display the math there.