jimmySixDOF

6 ms·

61.
by jimmySixDOF 1y ago
you can read the labels this (-y) uses modernBERT and even has an eval comparison to the (-ie) in it's GitHub so you can see the improvement as tested -- although if you want to do vanilla rules based chinking for whatever reason your
62.
by jimmySixDOF 1y ago
I did look into DataRobot's Syftr which points at the same problem but is a lot heavier I definitely like that the approach you take is at least easy to get a basic version up and can start checking the results right away!
63.
by jimmySixDOF 1y ago
I had my hopes on this project RawDog using local smol sized LLMs but it hasn't been updated in a while. I feel like all this should be running easily in the background nowadays. https://github.com/AbanteAI/rawdog
64.
by jimmySixDOF 1y ago
this kind of DSpy-GEPA self improvement loop keeps popping up and adding a few points but the cost (API and wall clock)also means you use this where a repeatable task/prompt/context needs optimizing and you can afford to find bett
65.
by jimmySixDOF 1y ago
I liked how Karpathy explained part of this problem as "silent collapse" in his recent Dwarkesh podcast. Meaning the models tend to fall into a local minima situation of using a few output wording templates for a large number of s
66.
by jimmySixDOF 1y ago
Also why the sudden interest? Amazon Alexa snips have been used before in court/investigation and this is not new. But makes me wonder about what happens when you are dealing with summaries of summaries of long gone tokens. Is that
67.
by jimmySixDOF 1y ago
There is Needle.Tools for porting Unity projects to WebGL/3js
68.
by jimmySixDOF 1y ago
WebXR is massively underrated
69.
by jimmySixDOF 1y ago
Agreed. These days setups imho aren't vanilla origination and termination VoIP scratch card traffic it's more likely a distributed bot farm obfuscation as a service provider. I have seen commercially available sim bank gateways t
70.
by jimmySixDOF 1y ago
Docling from IBM and Markitdown from Microsoft are reasonably reliable if you didn't try them also take the extra step to get image summaries in plain text from a VLM it's useful of you want to feed final results to an LLM later.
71.
by jimmySixDOF 1y ago
"No items" and "No featured items"
72.
by jimmySixDOF 1y ago
I'd be interested if you did any comparison testing to the langchain project which was, at least a month ago, the top open source approach https://huggingface.co/spaces/Ayanami0730/DeepResearch-Leade...
73.
by jimmySixDOF 1y ago
Will this protection extend to automobile companies ? Mobile Apps ? Mobile OSs ? I have lost track of the number of leakage points for location data into the tarball of databrokers.
74.
by jimmySixDOF 1y ago
Android has gone through epochs of customization option the best of which for me was 'Theamer' phase, a long gone one-click to customize the whole UI app and lot of creative people making things. Much much better than todays root
75.
by jimmySixDOF 1y ago
unrelated but the default power supply for Central Office telecom equipment was always -48V DC (and 23" vs 19" racks)
76.
by jimmySixDOF 1y ago
HCF came up in the recent IEEE Hot Interconnect Microsoft talked a few min about deploying in their HPC datacenters for AI latency reduction important to All Reduce/Gather operations typically in rings that need to converge and where s
77.
by jimmySixDOF 1y ago
While discussing Google Genie v3 and AndroidXR, Bilawal Sidhu said : "to create an even faster, lower latency pipeline to go from like 24 fps to like 100 fps. I could see that being more of an engineering problem than a research one at
78.
by jimmySixDOF 1y ago
IEEE Hot Interconnects just wrapped up and they discussed latency performance tuning for Ultra Ethernet where it looks smooth on 2- or 5- sec view but at 100ms you see the obvious frame burst effects. If you don't match your profiling
79.
by jimmySixDOF 1y ago
aaaannnnddddd ... it's happening: https://x.com/bilawalsidhu/status/1954229425199034753
80.
by jimmySixDOF 1y ago
I spent some time contemplating the analog of a shower or a bath and generally washing oneself and how that would look in a zero gravity based environment. My favorite idea is "water" injected into a contained environment to produ
81.
by jimmySixDOF 1y ago
What gets me is the egocentric perspective it has naturally produced from its training data, where you have the perception of a 3D 6 degrees of freedom world space around you. Once it's running at 90 frames per second and working in a
82.
by jimmySixDOF 1y ago
there is an active WebXR discord server you might get some feedback there too (I posted your project on cool finds channel btw ;) : https://discord.gg/xGKhKhSrDX
83.
by jimmySixDOF 1y ago
What about WebXR ? Also RP1 have a Metaverse Browser they are working on (see https://rp1.com/ ) could be interesting to discuss with them
84.
Cchistory: Tracking Claude Code System Prompt and Tool Changes(mariozechner.at)
2 points by jimmySixDOF 1y ago|0 comments
85.
by jimmySixDOF 1y ago
It's also not intended to be frozen in place and some allowance should be made for refinement with out scope breaking (fine but falls under change control) -- the level of detail here is important so known unknowns are surfaced not sol
86.
by jimmySixDOF 1y ago
tldraw is also something very easy to use but surprisingly powerful and I know they are compared a lot to Excalidraw, not so sure about draw.io - but could be worth adding to your list. It is a very extendable library with some fun genai h
87.
by jimmySixDOF 1y ago
there is a whole family of filetype glTF extension standards for vehicles so a completed asset will be rendered effectively across different game/world/ engines and coordinate the physics etc ... [1] is an example : [1] https:&#x
88.
by jimmySixDOF 1y ago
Vasturiano is a master what he does to combine D3 and ThreeJS is a huge decomplexifier everything can run in a single html file so simple so elegant so useful.
89.
by jimmySixDOF 1y ago
Little sparse on the documentation side can't tell at a glance if there is a 1:1 hyperperameter tuneability or if this is an opinionated single path locked soft fpga eval-hacking kind of thing. EDIT: -- Ok, it's legit, here is an
90.
by jimmySixDOF 1y ago
Modular says Max/Mojo will change this and make refactoring between different vendors (and different lines of the same vendor) less of a showstopper but tbd for now

More ›