michael

[Tibo @thsottiaux]

Old news actually from a bunch of days ago, but crossed that 15M. Enjoy a nice reset everyone. Landing in the next hour or so, go /fast.

[michael @michaelsCVR]

@aliasaria Please for the love of god give me a few hours with this. For those of us who had to deal with Claude lying or trashing CV projects for 5 months

[michael @michaelsCVR]

Any good YT channels in ML/AI? So much ai slop / reposting shorts

[Leading Report @LeadingReport]

Dad poses as his 11-year-old daughter on TikTok to lure her alleged rap*st to his home, then sh—ts him, per NYP.

[michael @michaelsCVR]

@ChangerHypnosis @fractalcounty When ai reaches that point I promise you, nobody will be safe in terms of skills. Skilled labor / trades obv different story for the time being. However ive found many ai users love projects that involve hands on work around house weirdly enough

[michael @michaelsCVR]

@IsThisA3DModel @fractalcounty errors are acceptable. This becomes largely irrelevant as AI advances, because it can repeat its output in an endless cycle until it reaches the actual solution or problem. That cycle has simply grown faster overtime

[michael @michaelsCVR]

@Vinocontonica @buccocapital What is the most difficult part - obv multiple things that all intertwine with one another. But if you had to list 1 or 2?

[michael @michaelsCVR]

@PeregrineGroupX @buccocapital I forget the exact quote but Zuckerberg answered something similar.. he was always being told about MySpace or other major competitors that could’ve destroyed Facebook easily at the start.

company has thousands of workers and endless $ is a bottleneck for true change.

[michael @michaelsCVR]

@AnkurGoel @buccocapital Hire some kid who is good with ai. Not sure how you’d find one, it’s definitely interesting problem that will become popular for some in the future. I’d be interested in just knowing what exactly you like about ERP & don’t like

[michael @michaelsCVR]

@bschne It really is just a tool that amplifies the imagination. Those with complex ideas that are changing rapidly (adhd type people) most likely benefit massively with ai.

[michael @michaelsCVR]

@LLMJunky For sure they pushed this hard realizing almost all the m1 MacBooks till now so many people aren’t upgrading. Apple seriously owns the laptop game since the M1 Pro. Such an amazing piece of tech. Never sold it, runs like its new still lmao

[michael @michaelsCVR]

@koomen Pete I work directly with largest east coast shipping port on site 2-4 days a week. Handle roughly 200-500 truck turns per day (season varies) and offload/loadback container ships. 7 STS cranes etc.

Different areas im sure but im sure bottlenecks are very similar

[michael @michaelsCVR]

@awilkinson @sama Claude code feels faster because It literally never thinks. Fable 5 opus 5 all of them are actual BS overhyped models. Sol ultra is the first model in a long time that I’ve experienced that for sure thinks and reasons

[michael @michaelsCVR]

@chamath What loop are you all in right now? Surely this must be because you are running frontier models with unlimited api / no $ limit. I have almost a pure certain degradation in quality of code/reasoning from all frontier models over the last 4 months

[michael @michaelsCVR]

@the_smart_ape @Trezor It’s literally an inside job. All these BS crypto draining is devs sitting on the knowledge for a few years then running the exploit. Nothing else, happens so often I don’t understand how users effected aren’t pitchforking the ceo’s house

[michael @michaelsCVR]

@SStricklandMMA Yet highest customer retention rate ever. Greatest piece of luxury ever created simply because it has the best tech. Tech= new luxury

[michael @michaelsCVR]

@yacineMTB That’s what you think. Google is 50 steps ahead, hiding behind all frontier labs. It’s funny how surface level some of you think.

[michael @michaelsCVR]

@chetaslua Had the same hype for fable then here we are today complete dog shit that excels in copying websites and making ai slop designs. Frontier labs purposely giving bad reasoning to extract training data to justify the subscription prices

[michael @michaelsCVR]

@AnthropicAI It’s literally an NLP with semantic matching. If this is working correctly then you guys aren’t using it at all or purposely not using it