userabsent.

My bio

[userabsent. @userabsenttoday]

@binarybits @worriedoptimist For the scenarios it is not relevant if some specialist AIs exist that is even better than generic AI. If generic AIs beat most experts (as already today domain-depending), it is sufficient due to "spin up 10000 experts that work 24/7 at 100x human speed" nature

[userabsent. @userabsenttoday]

@MSMInsider @SoWhy @NateSilver538 It is exactly the VAR who decides in this case if it is a serious incident or not. Better start learning about the game. Or just dont comment.

[userabsent. @userabsenttoday]

@D00kNuk3m @_mm85 you eat cow meat right? why would you eat that?

[userabsent. @userabsenttoday]

@benspringwater They could do so much better. No feature for a playlist from my music of <genre> with <mood>. Handling of classical music is atrocious, many features are not designed with it in mind. Podcasts dont hold a candle to dedicated apps. They are just not passionate about their product

[userabsent. @userabsenttoday]

@zeynep Seems you have not been using LLMs to build personal tools to improve productivity in certain tasks 2-10x (this holds for any knowledge worker). Now you can argue that improving productivity does not lead to replacing humans, but that would be actually worth discussing.

[userabsent. @userabsenttoday]

@NateSilver538 @ishaantharoor What are you guys doing for three days to fill out a bracket !? It takes 20 minutes?

[userabsent. @userabsenttoday]

@paolino skill issue, thats not sth to post about... next time ask an LLM

[userabsent. @userabsenttoday]

@investinguab Please tell your Claude not to write in AI tropes

[userabsent. @userabsenttoday]

@TheZvi @ai_ops_lead But the experiment would very likely prove you wrong? It is clear by now Fable is significantly more capable than Opus/GPT5.5 . So if "Fix this code" works to identify vulnerabilities (and it seems yes) than obviously Fable will find more of them? Seems crazy but logic is sound

[userabsent. @userabsenttoday]

@kallistosiii @favelaoverlord are you kidding, you are posting an article "Submitted on 5 Feb 2024 (v1), last revised 16 Mar 2025"

[userabsent. @userabsenttoday]

@krakenfx No it actually doesnt work. Even though SEPA Instant was used, the money was not available until 24h later. Disappointing

[userabsent. @userabsenttoday]

@MatthewJBar How can you calibrate on someones assessment of AI in 2016 ? Demis assessment was far off the mark too. How was yours at the time?

[userabsent. @userabsenttoday]

@Noahpinion Cheers to your father ! In the honor of hi story, how about you start researching topics more thoughtfully before posting about it

[userabsent. @userabsenttoday]

@slatestarcodex @souljagoyteller I cant answer that, but if you have ever had Noah write sth in your own area of expertise, you understand

[userabsent. @userabsenttoday]

@emollick These "Thinking steps" have nothing to do with the actual reasoning, they are generated separately to show "something". Thats why they always sound generic and repetitive.

[userabsent. @userabsenttoday]

@Jediwolf @SHL0MS I am sorry but are you that plain that you dont realize most answers are AI bots.

[userabsent. @userabsenttoday]

@joelgombiner @Conrguy @zeynep Yes the experts should tell you the risks, the likely trajectory, spelling out the uncertainties. All in a way that can be understood without a biology PhD. But no the experts shouldnt tell you what emotional reaction to have. Facts only

[userabsent. @userabsenttoday]

@DougJBalloon In no way would I defend RFK jr, but put simply, fluoride in the toothpaste is 100x more effective than fluoride in the water. So essentially fluoridized water is somewhat useful for preventing cavities in people that dont brush. For those that do, it has no effect.

[userabsent. @userabsenttoday]

@MrnllMtt @SereneInvesting @atc6955 Have you ever worked in an enterprise? Your expectation is silly

[userabsent. @userabsenttoday]

@MrnllMtt If a technology were invented that will make everyone (or a subgroup) 20% more productive, but right now doesnt yet, what would you expect to see a few months after the invention? In terms of Capex, and in terms of ROI? And no chatbots isnt the invention, its coding agents