1024core

20 ms·

271.
by 1024core 2y ago
MEGALODON builds on the research team's previous model, MEGA (exponential moving average with gated attention) I like how the team gingerly avoided calling the previous model "MAGA" ... :-D
272.
by 1024core 2y ago
Why does the Wiki page say "citation needed" for calling her a "computer scientist, electrical engineer"?
273.
by 1024core 2y ago
I like your ideas and would like to subscribe to your newsletter.
274.
by 1024core 2y ago
> The doc comment at the top of the .py file is sufficiently descriptive Which is the purpose of these doc comments. If you have the time to gripe on HN, you have the time to click on the link and do some reading. The "Usage" s
275.
by 1024core 2y ago
How did they get her number, to message her?
276.
by 1024core 2y ago
But the balloons stay up for weeks and float around; doesn't that cause problems?
277.
by 1024core 2y ago
How come their weather balloons don't run afoul of local flight regulations?
278.
by 1024core 2y ago
Pigeon towers seem to be making a comeback in India (and apparently they were found in ancient Persia too). I was visiting India last year and came across a colorful tower, about 60 feet tall and teeming with pigeons. There the aim is to pr
279.
by 1024core 2y ago
Any tools to replicate @oca.computer's work? Once we have the 1000-dim vector embeddings I can make the rest work. Not sure how to go from 20-word span to a 1000-dim vector embedding.
280.
by 1024core 2y ago
You will find similar pockets of responsive police in areas like Piedmont (in Oakland), Atherton, Hillsborough, etc.
281.
by 1024core 2y ago
Same of what to me?
282.
by 1024core 2y ago
You mean, like this? https://www.nbcnews.com/video/video-shows-moment-tesla-nearl...
283.
by 1024core 2y ago
I wouldn't call it a "real business" just yet. I've heard they do 50,000 rides per week in SF, LA, Phoenix combined. Assuming they make $20/ride, that's still $1M/week, or $52M/year. I'm sure the
284.
by 1024core 2y ago
> What's "open" about OpenAI? It's "open" for everyone to use, for the right fee.
285.
by 1024core 2y ago
You seem to have an axe to grind with @sama. This deep-rooted bias does not make for an honest discourse; as you are just expressing your opinions. If you have some facts to back up your claims, please put them out. Otherwise, I would urge
286.
by 1024core 2y ago
Even Jack couldn't handle being CEO of Twitter and Square. It's just not easy to do.
287.
by 1024core 2y ago
> As staff threatened to quit and speculation swirled that Microsoft may poach Altman's team from OpenAI and hire him directly, the company's board brought back Altman as CEO less than a week later. To me, this is an important
288.
by 1024core 2y ago
Thank you, from an appreciative reader!
289.
by 1024core 2y ago
Clearly there's room for innovation here. I'm sure someone (maybe a woman inventor, since they have first-hand experience with this stuff) will come up with a radical new design for a bike seat that works for them?
290.
by 1024core 2y ago
So GG Capital paid $2.1B for Red Lobster, and sold the land for $1.5B. They still owe $600M on their purchase! Did they sell the remnants to Thai Union for > $600M? Only then this whole thing makes sense.
291.
by 1024core 2y ago
Don't worry, AI and LLMs will save us.
292.
by 1024core 2y ago
RAIT (Redundant Array of Inexpensive Telephoto lenses)?
293.
by 1024core 2y ago
As Altman said (paraphrasing): GPT-4 is the _worst_ model you will ever have to deal with in your life (or something to that effect).
294.
by 1024core 2y ago
Thank you!
295.
by 1024core 2y ago
> For total water we can use +/ (“sum over”), resulting in +/{((|\x)&||\|x)-x}. And they say Perl is bad.
296.
by 1024core 2y ago
Reminds me of this story I heard about Yahoo. Their ads server had a memory leak and it would OOM after something like 10000 requests. Their solution: restart the server after 8000 requests. This worked for a year or two. And then it starte
297.
by 1024core 2y ago
Everybody loves to get outraged over every little thing nowadays.
298.
by 1024core 2y ago
IIRC his fund averaged around 30% gains per year, every year, over 30 years. (I'm going from memory here, too lazy to look it up). That is just such an unbelievable performance number.
299.
by 1024core 2y ago
> capable of up to 38 trillion operations per second Assuming these are BF16 ops, by comparison, an H100 from NVIDIA will do 1979 teraFLOPS BF16. So this "Neural Engine" from Apple is 50x slower than an H100 PCIe version.
300.
by 1024core 2y ago
If you have some time to kill, this site: https://www.officialdata.org/ca-property-tax/#37.76919241859... will show you your neighborhood's property taxes (Cali only, I think). You will find neighborhoods where on

More ›