Seth Pratt

what’s sota this week?

[Seth Pratt @sethprattsf]

@jun_song The 512gb refurbs are only $15k right now, I think you’re a bit high

[Seth Pratt @sethprattsf]

@0xSero interested

[Seth Pratt @sethprattsf]

@loktar00 not enough bandwidth

[Seth Pratt @sethprattsf]

@bubbleboi didn’t you just short micron in to earnings with $100k? 😂😂

[Seth Pratt @sethprattsf]

@MacRumors 512gb max, $18k msrp

[Seth Pratt @sethprattsf]

@bridgemindai time to buy more GPUs

[Seth Pratt @sethprattsf]

@AutismCapital @TheWuhanClan HomePods are terrible products that randomly play music until they die which happens within 2-3 years. Do not buy.

[Seth Pratt @sethprattsf]

@AlexFinn You should get rtx6000s not 5090s. The 5090 doesn’t have enough vram it is very limiting.

[Seth Pratt @sethprattsf]

@kimmonismus Bill Gates who was tied to Epstein in the Epstein files as a potential architect of Covid in order to profit off of mRNA vaccines?

[Seth Pratt @sethprattsf]

@pitdesi yes let’s blast ourselves with harmful led blue light

[sridhar @RamaswmySridhar]

Early results from the @snowflake's coco team on GLM-5.2 vs Opus-4.7 on dbt-bench — what the trajectories actually show 🧵

[Seth Pratt @sethprattsf]

@AnjneyMidha It’s called going to the gym and doong physical therapy. 😂

[Mat Velloso @matvelloso]

All day using GLM 5.2. Didn't miss much. First open model that passes the bar as a daily driver. Things are not going to be the same.

Damn, now I want to buy some serious hardware.

[Seth Pratt @sethprattsf]

@tonyGewrit because it is a cool lab experiment that has no concurrency and dramatically reduced quality versus the foundation model itself served over GPUs in a cluster?need 600 GPUs to serve 20 people at 30 tokens per second

[Seth Pratt @sethprattsf]

@oleksoleksoleks bullshit. I’ve rented clusters they are not cheaper than APIs in 95% of cases

[Seth Pratt @sethprattsf]

@TakeThiamine severely tore my hamstring a few days after starting on these. knee problems to this day 6 years later.

[Seth Pratt @sethprattsf]

@0xFelix1989 I’m already working on this. :)

[Seth Pratt @sethprattsf]

@banteg have to think about the future, what will this be able to run a year from now? fully abliterated within your control. no one training on your outputs so that other users can copy your novel contributions in one shot on the next release.

[Seth Pratt @sethprattsf]

@banteg thinking it will get 20t/s decode on sparks is generous

[Seth Pratt @sethprattsf]

@Takuiten also 0 concurrency.