reg

21

[reg @regietst]

U have free will

[reg @regietst]

This is a huge bet on open source. Maybe their plan is to become nvidia for consumers. Very interesting.

[AppleTrack @appltrack]

Apple's M7 Ultra chip coming in 2029 is rumored to support 1.5TB of RAM.

This would make the processor much more capable for on-device AI.

[reg @regietst]

@AnikaSomaia Walter isaacson steve jobs bio IMO

[reg @regietst]

@chamath Disagree. I believe these are all valid answers but like business you lose playing too many games. I think the key is optimizing for feasibility, which in this case means picking as many of these as you have bandwidth for.

[reg @regietst]

I cant fathom how AGI is on the horizon and yet search fails to work for 99% of consumer software products

[reg @regietst]

Bookmarks is borderline useless because of this @nikitabier

[reg @regietst]

@khushiirl Linkedin is x lite

[reg @regietst]

Many such cases

[Adam Rankin @rankintweets]

Every 3 years a new water bottle brand completely oneshots middle class white women like clockwork

[reg @regietst]

@sama https://t.co/VfsunIjb2P

[AFeldmanMMA @afeldMMA]

Conor McGregor speaking to Dana White and the media after injuring his knee in his first fight with Max Holloway.

Then rewatched exactly how it happened.

Thirteen years passed.

Different chapter. Same opponent. A different kind of heartbreak.

Time has this strange way of bringing certain stories full circle.

Whatever your opinion is of last night, it’s hard not to feel for the way it ended and for all involved.

[danawhite @danawhite]

McGregor vs Holloway is LIVE TOMORROW on @paramountplus! #UFC329

[Underrated Tribal Chief @CombatUTC]

“The Notorious” Conor McGregor vs Dennis Siver | UFCFN: Mcgregor vs Siver

McGregor kept Siver on the end of his punches and kicks, stopping him on a one sided beatdown in the 2nd as widely expected #UFC #MMA

Requested by @F______Dolphin

[reg @regietst]

1) what

2) someone needs to be fired for this

[Justin Skycak @justinskycak]

The greatest breakthrough in the science of learning over the last century:

[Brian Roemmele @BrianRoemmele]

Meet 1985 New York City king of the Bike Messengers.

I ran into this guy. He was literally everywhere.

This is 24 year old Leo, David Leopold.

He was a legend.

[boris @boristane]

god tier ad

[leo 🐾 @synthwavedd]

🚨 SCOOP(s):

- GPT-5.6 will be the final model in the 5.x series. GPT-6 is slated to launch in about a month, earlier than expected, and possibly even later this month
- GPT-6 will be based on a new, significantly larger pretrain (versus the ~4T 5.5/5.6 'Spud' base)
- There is lots of excitement at OpenAI over this new base, which they believe will be much better able to compete with both Fable 5 and upcoming 5.1, targeting a similar release window. OpenAI initially intended to continue with Spud through GPT-6, but decided against it
- On the topic of Fable 5.1, it is in the late stages of the pipeline at Anthropic and a release is expected "in the coming weeks"
- On the other side of the globe, DeepSeek are preparing for an imminent launch of V4 GA, which seems likely to be on par with or better than GLM-5.2, and have begun work on a new, larger model that will compete with the upcoming 2.7T MiniMax Pro

[OpenAI @OpenAI]

GPT-5.6 Sol, along with Terra and Luna, will launch publicly this Thursday.

We’re expanding preview access globally now.

[Bleys Goodson @bleysg]

It is a 2 to 4T param model. They are serving it across 70-100 wafers. To get healthy serving characteristics, they are essentially putting at most one layer per wafer, and the model is in the ballpark of 70-90 layers.

There's a couple of different ways this could be served and model sizes implied by that. One is if they keep the heavy KV caches they've used before. Another is if they go with lighter KV cache designs more akin to DeepSeekV4 or Hybrid SSM models.

The fact that they've partnered with Cerebras and designed with the hardware in mind means they're much more likely to have gone the second route. That SRAM bandwidth is too precious for a heavy KV cache. As such, something like the below is the center of probability mass: 3T total, 150B active, 70 layers.