i hadn't of this podcast but, as a fan of intervals . icu, i'll check it out on my ride today.
08.03.2026 15:14 — 👍 0 🔁 0 💬 0 📌 0i hadn't of this podcast but, as a fan of intervals . icu, i'll check it out on my ride today.
08.03.2026 15:14 — 👍 0 🔁 0 💬 0 📌 0it sort of reminds me of the widespread racism issues reported at tesla and elon's entire response was "grow a thicker skin"
06.03.2026 17:13 — 👍 115 🔁 5 💬 3 📌 0a male tabby cat on a bright day, sitting alert on an old wooden deck with his front legs straight and his rear resting on his haunches. He's wearing a bright orange harness. his green eyes have extremely narrow pupils.
my cat demands to go outside lately to verify the security of our yard.
06.03.2026 16:29 — 👍 2 🔁 0 💬 1 📌 0
Is it 1997 again?
Can you bring back 1997?
where did you see that?
05.03.2026 18:45 — 👍 0 🔁 0 💬 0 📌 0Can they at least record it and stage it to be a little more dramatic? Then he could revive his catch phrase!
05.03.2026 18:19 — 👍 0 🔁 0 💬 0 📌 0That's hard to believe tbh
05.03.2026 18:17 — 👍 1 🔁 0 💬 0 📌 0What a wonderful way to look at it
04.03.2026 12:46 — 👍 2 🔁 0 💬 0 📌 0i'm specifically interested in the c1x modem and don't want the air.
03.03.2026 14:24 — 👍 0 🔁 0 💬 1 📌 0Claude works better than openai's models, at least
03.03.2026 13:30 — 👍 2 🔁 0 💬 1 📌 0
i never could get gemini to admit it didn't know something: it would just generate output imagining interface options that don't exist, etc. Claude did.
Gemini would provide 3 paragraphs of output (or 5) when 2 would do. Claude is more concise.
This is with sonnet 4.6.
I did as honest a comparison of Claude to Gemini this weekend as i could for my personal use.
I feel Claude is a fair bit better its ability to reasonably push back when you may be wrong, sometimes admit not having a good answer, not being overly verbose, and just getting things done.
i only use the office 365 calendar for work and almost exclusively through OWA and i'm shocked at how poorly it works in that interface. I would never choose teams or OWA.
03.03.2026 00:48 — 👍 0 🔁 0 💬 0 📌 0i'll hold on to my pixel 9 and try the iphone for at least 30 days before deciding which phone to keep. There are a lot of social benefits to having an iphone (sharing notes, group chats, etc) that don't translate as well from iphone to android.
03.03.2026 00:47 — 👍 0 🔁 0 💬 1 📌 0
after using google phones exclusively since the galaxy nexus, i'm going to try iphone with the 17e.
The lack of UWB is the only thing that i find disappointing. I'd trade that for magsafe. I think i prefer the c1x modem to the base iphone 17 modem as i expect it to be a little more energy efficient
love everything but that 25mph limit ;)
02.03.2026 14:58 — 👍 1 🔁 0 💬 0 📌 0that is the worst use of a language model and your time i could think of
01.03.2026 22:22 — 👍 1 🔁 0 💬 0 📌 0
That phone opened to messages but held against his ear reminds me of movies or shows where the phone is glowing on someone's cheek as they pretend to have a conversation.
When the phone app is open, it would use the proximity sensor to turn off the screen. It would be a bit distracting to me.
And if you speed up 30% of my workload perfectly, compressing it 100-to-1 or better, i still have 70% of my work where using a model is a wash or worse.
It feels handy-wavy or even made up.
So while i agree it can save massive amounts of time (~30 minutes uninterrupted, if you can get it, down to 5 minutes as one example) but there are so many areas where it is limited that it can only speed up maybe 20-30% of my total work load.
28.02.2026 22:19 — 👍 0 🔁 0 💬 1 📌 0but for a lot of operations i find it takes me longer in a lot of cases to work it through a language model compared to just reading documentation and solving it myself, because the models are so often generating hallucinations or not getting it quite right.
28.02.2026 22:19 — 👍 0 🔁 0 💬 1 📌 0
In my experience, they can be very helpful for some things, "generate a dockerfile that sets up these sets of packages, installs everything necessary for linux RDP server, and creates the following services ready to go with podman."
You could have something ready for testing in minutes.
I'd also question the metrics you're using to measure productivity - how can you estimate the productivity those 100 employees generate without language models.
28.02.2026 22:19 — 👍 0 🔁 0 💬 1 📌 0
i presume you're referring to companies using the products, not the companies creating and selling the products.
nvidia's demand is sky high because of companies buying their products, or planning to buy their products (coreweave, openai, oracle) in massive volumes with negative returns so far.
honda asimo robot on a blue background signaling 'i love you' with each hand
it makes sense because the tesla bot is basically a Temu version of the Honda Asimo
28.02.2026 22:08 — 👍 3 🔁 0 💬 1 📌 0Maybe it was meant to be a Piece Prize
28.02.2026 21:04 — 👍 1 🔁 0 💬 0 📌 0It may pop the Bitcoin bubble but it's still a lot of spending without being remotely close to profits for most of the companies in the area
28.02.2026 18:16 — 👍 0 🔁 0 💬 1 📌 0It may be overvalued but at least gold is real and also has uses.
28.02.2026 17:54 — 👍 1 🔁 0 💬 0 📌 0i like this because you were a huge elon/tesla fan from the beginning, and an early model 3 adopter and supporter so you showing this means something
25.02.2026 04:37 — 👍 0 🔁 0 💬 0 📌 0
new update
this time it crashed.