malchow

Well-run autolabs recursively training discipline-specific models are becoming very valuable assets. I write more about this here:

https://x.com/jmalchow/status/2067298271647904061

As time goes on, the value of the model will go down and the value of the tools will go up.

For those who may have forgotten, Mattermost is quite good these days: https://mattermost.com/

What's interesting is you might not want to see de novo AI-generated storytelling (slop factor), but you might really like the way AI can make a story crafted by humans more interactive.

You can add your car to your whole-house AC bus using the Enphase bi-di EVSE, releasing this year:

https://enphase.com/ev-chargers/bidirectional

Satellite signals are just weak RF signals and can be disrupted easily. There is nothing 'hardened' about them. It's funny that people think Starlink or any of its many incipient competitors are any different.

This is an incredible foretaste of what AI can enable in gaming. Not replacing humans (the creators here are former leaders from Minecraft), but rather simply unlocking more fun gameplay by offering creativity, humor, and branched storytelling customized to the player.

I have a Tesla Model X and the front door and rear door aren't even latitudinally aligned.

In addition to bi-facials starting to work quite well, HNers may be interested in a rising class of ultra-low-mass material that has come out of work at Stanford and Intel in transition-metal dichalcogenides (TMDs) (e.g. for MoS₂, WS₂, WSe₂, etc.).

It turns out these enable a very high specific-power PV cell that adds another even more attractive production curve behind what is happening in vertical bi-facials. See e.g.:

https://ee.stanford.edu/frederick-nitta-koosha-nassiri-nazif...

https://www.arinna.xyz/

Cells are subject to aggressive binning – akin to the early days of MOS process. And specs and process capability adhesion in the industry are not as "serious" as they are in the modern, U.S.-led semiconductor industry.

Source: I work on 100% Si anode batteries constructed in part with a litho-derived laser process at Enovix.

Practically, shorter path silicon does nice things for power budget for a few minutes, that then gets totally obliterated by feature builders wanting to do cooler things!

100% active Si material in the anode. Not 5% doping of graphite.

Sneak peek of the production process: https://www.youtube.com/watch?v=LYZZa39nvDM

Amazing.

Not for cars yet, but Enovix has short-ion-path batteries intrinsic to its lithrographed 3D design. Performance is quite something: https://www.enovix.com/products/

Arbitrary vehicle to home/home battery/grid connection is indeed coming in very short order.

https://enphase.com/ev-chargers/bidirectional

There are other products already available to do it (DCBel), and it can be hacked of course, but at the current moment everything comes with substantial corner case blind spots, mostly related to grid-forming/following switching and to the resilience of the power electronics.

You are interpreting the entire story from the perspective of the utility and ISO (which are distributors, neither a maker of energy nor a net consumer of it). They're Edison-era inventions, borne of the lumbering size, toxicity, noise, and poor capacity factor of mechanical engines.

Why don't you interpret it from the perspective of companies (which need to buy energy to do stuff) and consumers (who need to buy energy to be comfortable)? These are the actual buyers.

Then you'll see why solar and batteries win all, and utilities, with their badly potholed power highways with massive tolls to consumers, slowly reduce in importance.

The power grids (in the US) do not do a good job of keeping 60 Hz these days. More here:

https://enphase.com/download/calculating-ac-line-voltage-ris...

The more Enphase type inverters (https://enphase.com/download/iq8-and-iq8-microinverters-data...) that can be introduced to the grid, the closer we can get to stable voltages as in days of old.

You are exactly correct, and the purism represented by these Kafkaesque Wikipedians who "drive out fringe editors" will ultimately fail.