pillefitz

10 ms·

91.
by pillefitz 1y ago
A while back I built https://readboost.io/ but so far never advertised for it. The idea: Embed Q&A into your ePubs and optionally download Anki cards for your book. Probably gets buried here, but maybe someone finds som
92.
by pillefitz 1y ago
Working in management, OneNote is my second brain I couldn't do without. Pages of meeting notes everyday, documentation of internal processes etc., all of which I regularly look up.
93.
by pillefitz 1y ago
I was recently reading another book of Ballard while on the flight to London and, shortly after arriving, visited the Barbican. What a magical experience it was!
94.
by pillefitz 1y ago
At least FAANG and well-funded startups did a lot of overhiring. When thinking about stable jobs I have classically boring ones in mind, such as ones in utilities or defense
95.
by pillefitz 2y ago
They are useful to reach Arc-N+1
96.
by pillefitz 2y ago
"First they came for the socialists, and I did not speak out—because I was not a socialist. Then they came for the trade unionists, and I did not speak out—because I was not a trade unionist. Then they came for the Jews, and I did not
97.
by pillefitz 2y ago
Dear Americans, please do something. Protest, write letters to congressmen etc. We Germans went through all of this 90 years ago..
98.
by pillefitz 2y ago
A few counter points: Most developers prefer Claude 3.5 (or the slightly improved 3.7 since a few days back), which is a year old already. What was supposedly ChatGPT 5 will be called 4.5 instead, after two multi-month and hugely expensive
99.
by pillefitz 2y ago
In image processing at least, NN typically learn a Fourier or Wavelet representation in their first layers. Biggest benefit of applying a transformation beforehands is to reduce training time / obtain better generalization by "rem
100.
by pillefitz 2y ago
Could the default be misalignment, and by fine-tuning it on anything we're deviating from the local optimum we call alignment?
101.
by pillefitz 2y ago
You're not doing whatever oppressed group you're representing a favor by using weird, alienating pronouns.
102.
by pillefitz 2y ago
Looks cool! Any luck finding paying customers yet?
103.
by pillefitz 2y ago
Awesome, I did something similar a while ago. What's your motivation? F
104.
by pillefitz 2y ago
Why is that? GTD for me was the only system that stuck. Set it up once and have it synced across all devices (I use MS ToDo, as it's free, comes pre installed ony work computer and doesn't have many fancy features). The secret is
105.
by pillefitz 2y ago
I just use MS ToDo synced across all devices with GTD. My "Waiting For" bucket currently contains around 20 items, most of them with reminders to follow up on the status. No way I would keep on top of it without a structured syste
106.
by pillefitz 2y ago
Bigscreen Beta if you have a VR headset. Let's you e.g. sit in front of a fireplace with a friend and watch or listen to YouTube
107.
by pillefitz 2y ago
I rather spend life in a poor democracy than a rich technocracy which supports dictators around the globe.
108.
by pillefitz 2y ago
I'm increasingly happy we have these regulations that prevent us from being ruled by the likes of Musk.
109.
by pillefitz 2y ago
Eat your own dog food, i.e. use your own product
110.
by pillefitz 2y ago
If I can host the backend myself, does it imply it's not a commercial offering?
111.
by pillefitz 2y ago
Cutting subsidies would weaken Tesla's competitors (according to Musk)
112.
by pillefitz 2y ago
They were at least united and somewhat effective in commiting these atrocities.
113.
by pillefitz 2y ago
Does it matter if they mimicked it better than most humans?
114.
by pillefitz 2y ago
Why is that?
115.
by pillefitz 2y ago
Which is related to multistage/ hierarchical/coarse-to-fine optimization, which is a pretty good way to find the global optimum in many problem domains.
116.
by pillefitz 2y ago
It's been done already https://aipapersacademy.com/chain-of-continuous-thought/
117.
by pillefitz 2y ago
Now I'm intrigued. Why not post it publicly?
118.
by pillefitz 2y ago
Only difference is those startups don't rub it into your face they're database-based.
119.
by pillefitz 2y ago
Since most foundational models have been trained on illegally acquired books, this info should be already baked in.
120.
by pillefitz 2y ago
It all feels like an attempt to prevent replication to further tank the market. Not necessarily the technical details, but the reporting thereo, which was spammed all over Reddit and HN

More ›