Xmd5a

28 ms·

361.
by Xmd5a 2y ago
Yes, I have, and this is a bad idea (strawberry + mint). The problem is that the herbs become harder than the ice cream, creating an unpleasant sensation in the mouth. I had a similar issue with a toffee ice cream where I tried incorporatin
362.
by Xmd5a 2y ago
I started cooking ice cream this summer. I bought an ice cream machine that can yield 1L of ice cream in 30 minutes, and prepared about 30L, sometimes sourcing material from the garden (strawberries, raspberries, cherries, figs and blackber
363.
by Xmd5a 2y ago
Here's one: As relayed by Russian propaganda outlets: https://x.com/VigilantFox/status/1820724008637063664 Reuters fact check: https://www.reuters.com/fact-check/video-arrest-over-faceboo
364.
by Xmd5a 2y ago
>USSR was also terribly battered by WWII, and its leadership was not highly competent either The USSR moved all its industry eastward, as the German army advanced, waiting for the very last moment to do so. Quite an incredible feat that
365.
by Xmd5a 2y ago
- The Occupy Wall Street movement. - A COINTELPRO-inspired diversion undermines the cause: during demonstrations, individuals wishing to speak must wait in line, while women, minorities, and other groups are prioritized. - This method becom
366.
by Xmd5a 2y ago
The first plane to ever fly was in fact ignored by the general public for several years https://bigthink.com/the-past/wright-brothers-ignored/
367.
by Xmd5a 2y ago
No but they gave us GAI. The fact they flipped the frame problem(s) upside down is remarkable but not often discussed.
368.
by Xmd5a 2y ago
Search is the most if not only useful application of LLMs. In this space there is also Phind. It seems I use Perplexity as often as Google nowadays.
369.
by Xmd5a 2y ago
I think the main reason why diagrams are not used as often as they should to document code bases is that we developers have to face opposite yet complementary approaches: text based graph specifications (but then the layout's quality d
370.
by Xmd5a 2y ago
I tried TALA and am surprised to learn it embeds a machine learning model. I settled on ELK (very complex and time consuming to get right), before getting interested in constraint based graph layout (mostly for aligning nodes). Symmetry is
371.
by Xmd5a 2y ago
Indeed, but consider this situation: You have a collection of documents and want to extract the first n words because you're interested in the semantic content of the beginning of each doc. You use a LLM because why not. The LLM proces
372.
by Xmd5a 2y ago
https://pressbooks.openedmb.ca/wordandsentencestructures/cha... I believe what the LLM replies with is in fact correct. From the standpoint of a programmer or any other category of people that are attuned to some kind
373.
by Xmd5a 2y ago
Also, some interesting links I stumbled upon while exploring this space: Graphviz-like generic graph-drawing library. More options, more control. https://eclipse.dev/elk/ Experiments by the same team responsible for th
374.
by Xmd5a 2y ago
I only use diagramming tools with automatic layout algos to generate visualizations programmatically to debug things. For documentation purpose, I largely prefer to draw them by hand using OmniGraffle, because making graphs look nice using
375.
by Xmd5a 2y ago
I'm not into ML so I can't answer your question specifically, but it seems your circular google translate or jpeg comparisons are missing the essential element: self-evaluation https://x.com/search?q=synthetic%20da
376.
by Xmd5a 2y ago
>the concept of surprise is very cool Then you may be interested in Simplicity Theory: https://simplicitytheory.telecom-paris.fr/ Relevant situations are unexpected Relevant features generate compression A
377.
by Xmd5a 2y ago
>What stops them suffering the same entropy issues? Is there a reinforcement step with human feedback? It seems the situation changed with inference scaling/reasoning models (o1 and beyond). I found the tweet I was looking for, but
378.
by Xmd5a 2y ago
Related tweet by Sam Altman: https://x.com/sama/status/1875603249472139576 If you can find it in this thread, an OpenAI employee hinted at the idea they have reached "data singularity": using a model to
379.
by Xmd5a 2y ago
Thanks for the correction, I guess I was lured by yet another LLM confabulation
380.
by Xmd5a 2y ago
>【DIY】3D-Printed Visual-Guidance Surface-to-Air Rocket (Making) https://www.youtube.com/watch?v=UvcDwSmmxWs
381.
by Xmd5a 2y ago
I had to look up MCST: it means Model-Centric Software Tools, as opposed to autonomous agents. Devin is closer to a long-running process that you can interact with as it is processing tasks, whereas Cursor is closer to a function call: once
382.
by Xmd5a 2y ago
Indeed this is what I was aiming at, however the concern for (a semblance of) truth seems rooted in a view that locates meaning in what signs refer to. This view feels incomplete when faced with a dyadic model where the relationship between
383.
by Xmd5a 2y ago
In control theory, this is called dead-time: >Dead time is the delay from when a controller output (CO) signal is issued until when the measured process variable (PV) first begins to respond. https://controlguru.com/dead-
384.
by Xmd5a 2y ago
What's your take on LLMs ? I ask you to comment on any aspect, whatever you think is the most interesting from a semiotician's perspective.
385.
by Xmd5a 2y ago
I tried aider and had problems having it update code in existing files. Aider uses a search and replace pattern to update existing code. So you often end up with >>>SEARCH } >>>REPLACE }, {'m
386.
by Xmd5a 2y ago
I don't understand what point you're trying to make. Life never appeared on Venus or Mars as far as we know. The question I want to ask is: why is there almost no existential pollution problems in nature ? Why does the way civiliz
387.
by Xmd5a 2y ago
https://x.com/amnestyfrance/status/1864689106032214016 >You have become a shop for far-left ideologues who choose their causes based on party doctrine. Take not of the pro-Israel agenda that grafted itself on t
388.
by Xmd5a 2y ago
>"flattens the CO2 curve" There is no such thing as polluting substances, only an accumulation of those. The problem lies in the fact that our technological process graph has too many sinks leading to an accumulation of 'd
389.
by Xmd5a 2y ago
https://github.com/mit-han-lab/streaming-llm On a side node, and that's what led me to the link above, I wonder if it would be possible to chain N streaming LLMs in an agent workflow and get a final output stream
390.
by Xmd5a 2y ago
Says the man who has settled in the basement.

More ›