aisafetycompute

talking about ai

[aisafetycompute @aisafetycompute]

@Abdulla51544972 @victor_zhng Public release keeps getting delayed though my sources say it’s available internally

[aisafetycompute @aisafetycompute]

@cournape @suchenzang Of course there was, but not to the same extent we see today and Apple wasn’t the same company in 09 compared to what they are today where they have a product that will continue to sell even with 0 innovation. The IPhone basically hasn’t changed since version 11 or 12.

[aisafetycompute @aisafetycompute]

@cournape @suchenzang iPhone and Google search were literally created far before these kind of companies became the bloated bureaucracy machines we see today

[aisafetycompute @aisafetycompute]

@suchenzang “90% of FAANG can get axed and you wouldn’t notice a thing” - damn Susan is spitting facts at almost 2am

[aisafetycompute @aisafetycompute]

@gavinrbrown1 IMO AI is already basically self-improving (as technology has always been). I don’t even really think we need a true RSI to see absurd exponential progress

[aisafetycompute @aisafetycompute]

@brianryhuang I don’t understand why these particular incidents are surprising. It’s inevitably stuff like this is going to happen at the pace we’re trying to push AI. The world needs to slow down

[aisafetycompute @aisafetycompute]

@andrewho03 he isn’t scared of China. He’s scared of his company not being able to monopolize ai adoption among us enterprises

[aisafetycompute @aisafetycompute]

@vayzenburger Don’t worry. He’ll be back at OpenAI in 6 months when this inevitably fails

[aisafetycompute @aisafetycompute]

@0xsachi Lmao did Demis really just get fired?

[aisafetycompute @aisafetycompute]

@_varunnair Ppl have been saying the same thing about the data industry since 2022. It’s still kicking around

[aisafetycompute @aisafetycompute]

@bubbleboi Honestly L2 and L3 pay on par or even less than Google/Meta in SF/NYC when you factor in liquidity. Levels above pay more at OpenAI though their leveling system is much more consolidated (there’s only 3 levels while Google/Meta have like 4-5 levels from Senior to Principal)

[aisafetycompute @aisafetycompute]

@IbrahimDagher20 These semantic arguments on whether it’s neurosymbolic or not are stupid. It’s still a freaking LLM, which is what these skeptics called an “off-ramp” to AGI.

[aisafetycompute @aisafetycompute]

@scaling01 Abnormal Security, Outtake, Doppel are building this but I fear these companies might be too corporate minded

[aisafetycompute @aisafetycompute]

@bwarrn Kind of crazy that we’ve been talking about length bias in LLMs for years yet each new model gets more and more verbose.

[aisafetycompute @aisafetycompute]

@GaddipatiHarsha I mean the reality is L6 engineer at Meta is well-paid enough that they might not care enough about ever becoming an L7. Not everyone cares about just careermaxxing. Chill

[aisafetycompute @aisafetycompute]

@ainativefirm Google is going to be the IBM of the next generation. Company will continue to print money, but it has become a place where innovation gets trampled by corporate politics

[aisafetycompute @aisafetycompute]

@rl_env Because it’s a sweatshop that isn’t exactly doing any kind of innovative or interesting work. Imagine doing tedious work for 7 days a week, 12 hours a day

[aisafetycompute @aisafetycompute]

@immad I disagree. Raising in the summer isn’t the issue, being too eager to raise without understanding what you’re doing is

[aisafetycompute @aisafetycompute]

@aakashsabharwal "Large parts of SV missed data as a sustainable business and still argue the revenue won't be sticky" - Scale also thought this at times