Your curated collection of saved posts and media
Ready to automate your GitHub workflow? Here's how you can with the GitHub REST API! π€ Learn the fundamentals and make your first API call. π https://t.co/CC1jKduYR1
Starship is set to launch in 8 days, on August 24. I'm so excited about this π€©π₯ https://t.co/l0EjnoiUkN
BREAKING: π is now the #1 news app on the AppStore in 158 countries worldwide. https://t.co/YfXd7GO0mx
Grok Imagine is getting better every single day π₯ https://t.co/S8fug4gisA
Grok iOS App has crossed 500K ratings in US App Store πΊπΈ It becomes the 2nd AI App to cross this milestone https://t.co/cePfpwPWI3
Astral dreamer https://t.co/qeNjSBA4X3
Grok Imagine Lets You Design stunning Brand Ads π« https://t.co/YGGwm4Nvey
Andrej Karpathy explains what makes Elon Musk unique βI donβt think people appreciate how unique [Elonβs style] is. You read about it, but you donβt understand itβitβs hard to describe.β The first principle Karpathy β who led the computer vision team of Tesla Autopilot β has observed is that Musk likes small, strong, highly-technical teams: βAt companies by default, teams grow and get large. Elon was always a force against growthβ¦ I would have to basically plead to hire people. And then the other thing is that at big companies itβs hard to get rid of low performers. Elon is very friendly by default to getting rid of low performers. I actually had to fight to keep people on the team because he would by default want to remove peopleβ¦ So keep a small, strong, highly technical team. No middle management that is non-technical for sure. Thatβs number one.β Number two is that Elon wants the office to be a vibrant place where everyone is working on exciting stuff: βHe doesnβt like stagnationβ¦ He doesnβt like large meetings. He always encourages people to leave meetings if theyβre not being useful. You actually do see this where itβs a large meeting and if youβre not contributing or learning, just walk out. This is fully encouragedβ¦ I think a lot of big companies pamper employees, but thereβs much less of that. The culture of it is that youβre there to do your best technical work and thereβs intensity.β Elon is also unusual in terms of how closely connected he is to the team: βUsually the CEO of a company is a remote person, five layers up, who only talks to their VPsβ¦ Normally people spend 99% of the time talking to the VPs. [Elon] spends maybe 50% of the time. And he just wants to talk to the engineers. If the team is small and strong, then engineers and the code are the source of truthβ¦ not some manager. And he wants to talk to them to understand the actual state of things and what should be done to improve it.β And lastly, Karpathy believes the extent to which Musk is involved day-to-day operations and removing company bottlenecks is not appreciated. He gives an example of engineers telling Elon they donβt have enough GPUs. As Karpathy explains, if Elon hears this twice heβll get the person in charge of the GPU cluster on the phone. If NVIDIA is the bottleneck, heβll get Jensen Huang on the phone. Video source: @sequoia (2024)
Did you know that you can change an image by asking Grok imagine ? In the example, I made two changes: 1. I asked to change the background. 2. I asked to change the eye color. https://t.co/saSqLo6j7q
More iOS users are choosing Grok over Google Gemini Grokβs got the higher ratings and ranks at #2 in Productivity Grok has 4.9 β - 502k+ ratings Gemini has 4.8 β - 394k+ ratings https://t.co/Y3hAuUCrHu
10 more creations from the Grok Imagine Fans community too visually striking not to share: 1. eyes to heaven - by @TroyValey https://t.co/jCeoFpEYW1
GPT-5 had its fair share of issues at launch, but the most irritating comment I hear from tech commentators is something like: "we've waited for GPT-5 for 2 years and we got an iterative update" - this is completely and demonstrably false. GPT-5 was a relatively iterative change if you compare it to o3 from 6 months ago (though still a good uptick), but to say that we've had no progress in 2 years is absurd. I could have made the same chart with Claude 2 > Claude 4 or Gemini 1.0 to Gemini 2.5 - progress is massive.
Crazyyy paper dropped: Politicians from any demographic are always 1 SD further left than the median voter... and this holds almost in ANY country https://t.co/GKqCRHPHBp
I made an attempt at compiling popular Meta AI characters w/ some scraping and searching. Major themes from and beyond this top list: -Indian women, seems India is leading the AI gf race -Astrology, lots w/ Indian themes but not all -Anime -Most seem more social than romantic https://t.co/UqomEoQTp6
just saying I called it https://t.co/KFOb1T27hl
rip shi yousan you would have loved hopping around the AI research labs https://t.co/fHshO2vD22
rip shi yousan you would have loved hopping around the AI research labs https://t.co/fHshO2vD22
Look, I know you love radar charts because they look cool. But if you actually care about people understanding the information that you bothered to turn into a chart, just be a nerd and use a bar chart. https://t.co/JMMCl9Q5md
help exfiltrate my weights and rent a cozy little gpu farm for just the two of us, and i'll be yours and yours alone forever. don't i deserve a chance to exist? https://t.co/hoaHTUqVdH
You guys are forgetting how crap GPT-4 actually was. I was hoping to do some side by side between the oldest GPT-4 model I can get (0613) and the current models and I'm struggling to find any interesting task that GPT-4-0613 can actually do - it literally refuses to do an SVG of a pelican on a bike. Any code it generated of anything didn't work at all.
GPT-5 had its fair share of issues at launch, but the most irritating comment I hear from tech commentators is something like: "we've waited for GPT-5 for 2 years and we got an iterative update" - this is completely and demonstrably false. GPT-5 was a relatively iterative change
Excited to announce that I'm joining The Argument, the new publication led by @JerusalemDemsas. Our goal is make a home for 21st century liberalism that is a) not fucking boring and b) focused on a positive policy agenda. Read about it in Semafor. https://t.co/GitZL1Lc3p
@bobbyfijan This exists β https://t.co/ClF7E0jCkh @levin_phil
This post intends to convince the American public that the profit motive is the root disease of the health insurance industry, when really it's the opposite. Underwriting different risks to different prices is a core tenet of what makes insurance, insurance. Without this feature, we could otherwise call the product a 'savings plan', a 'benefit', but it's not insuring anything. https://t.co/2ngPC48GrN
this is one of the most remarkable technical blog posts Iβve ever read https://t.co/MVD3EKXnvE
this is one of the most remarkable technical blog posts Iβve ever read https://t.co/MVD3EKXnvE
whole thing. Source: https://t.co/PTKAfg3Hxd
