dist-epoch

6 ms·

31.
by dist-epoch 1mo ago
You can use regular voice dictation to talk directly to the Astra model, using an app like Handy. Then you have zero message passing overhead. I do that when I know exactly what I want, and use ChatGPT voice when I need to explore the solut
32.
by dist-epoch 1mo ago
Maybe a one line change takes 2 min because it needs to run the full test suite and so on. I have special instructions in my AGENTS.md to bypass running the full QA suite for small contained changes, and to run a targeted one instead. And i
33.
by dist-epoch 1mo ago
You will have a good understanding of a system which is 20 times smaller in size/features. For a todo app that is fine. But for most projects your competitor will leave you in the dust with the speed at which they add features. Same st
34.
by dist-epoch 1mo ago
I used it in a similar way, and once you get it going, it's much faster, since it can reuse the background agent which now has a primed context. I got 30 sec turn around times. And you can talk about the next change while the previous
35.
by dist-epoch 1mo ago
`codex -p` is the subscription equivalent. or ACP if you want to be fancy
36.
by dist-epoch 1mo ago
You can give the agent a tool (or bash script) which waits for events. Agent calls it and the tool sleeps until an event happens then returns it to the agent.
37.
by dist-epoch 1mo ago
Nobody wrote 13 mil lines proofs before. I'm pretty sure you can make Lean at least 10 times faster if you unleash the agents on it. Somebody ported Doom to run entirely in the TypeScript TYPES (not code). It took 12 days to compile.
38.
by dist-epoch 1mo ago
Before transferring to life, it doesn't even transfer to a game like Poker.
39.
by dist-epoch 1mo ago
You could consider ZFS a VCS, and it can easily store multiple versions (snapshots) of Visual Studio. It's not impossible, there just isn't that much demand for it.
40.
by dist-epoch 1mo ago
One has nothing to do with the other. It was long predicted that math and software developments would be the first domain where AI was going to do major damage. If OpenAI and Anthropic didn't get into math result dick measuring, Intern
41.
by dist-epoch 1mo ago
There have been about 6-8 major math breakthroughs claimed by AI. Only for 2 of them there are public accusations about the training data.
42.
by dist-epoch 1mo ago
It's all marketing. Don't fall for it, it's the age old strategy "our product is extremely dangerous, so fear us". Like the tobacco companies saying all the time "our cigarettes are so dangerous, they cause can
43.
by dist-epoch 1mo ago
OpenAI says they just pointed the AI at the problem: > Regarding the level of human involvement on our end: although a group of people was involved in our efforts, we collectively had no research-level expertise in fluid dynamics and the
44.
by dist-epoch 1mo ago
This is the plot of 3 Body Problem, the Dark Forest. You need to hide yourself (the problem you are working on) or the super advanced aliens will obliterate you (start working on your problem) the moment they know you exist (rumors the prob
45.
by dist-epoch 1mo ago
Why not become a machine gun operator?
46.
by dist-epoch 1mo ago
> I’m using models, I’ve built two whole projects without looking at more than 1% of code, but I’ve tested the interface and went through data on disk to verify the assumptions. so, would you say "no worries man I know what I'm
47.
by dist-epoch 1mo ago
if you have a computer to run said local GLM models one
48.
by dist-epoch 1mo ago
You are conflating two concepts into one words, immigration and illegal immigration are not the same thing
49.
by dist-epoch 1mo ago
> with Pax Americana looking increasingly shaky. that is being replaced with Pax Silica https://www.state.gov/pax-silica
50.
by dist-epoch 1mo ago
You could use this model as another input into the country-scale models.
51.
by dist-epoch 1mo ago
VMware Workstation has been free for about 2 years now. https://blogs.vmware.com/cloud-foundation/2024/11/11/vmware-...
52.
by dist-epoch 1mo ago
having used zstd, it has terrible defaults, optimized for speed and low-memory. You need to change it's params (not just level and dict size) to get high performance. probably somebody should use a coding agent to do auto-research to o
53.
by dist-epoch 1mo ago
AI boosters: "look, AI is creating jobs, all of these doomers scaring you with job loses, don't listen to them" also AI boosters: "we are going to build a million robots, and the robots will build factories producing rob
54.
by dist-epoch 1mo ago
with zstd at level 16 with default params (dict size, ...). Serious compression starts at level 19 and with much higher dict sizes. how is this an honest benchmark: bzip3 ... 12178M memory zstd ... 687M memory
55.
by dist-epoch 1mo ago
The market says the problem is still there. An NVIDIA consumer GPU sells for 50+% or more than an equivalent AMD GPU. Because people are buying NVIDIA GPUs to run local models instead of AMD ones. I did the same thing, I paid 50% more to ge
56.
by dist-epoch 1mo ago
George Hotz in June 2023: > I have had direct contact with members of the AMD RTG team and I was disgusted to find that AMD doesn't even provide them with hardware to work on. The developer I was working with had to buy the GPU he w
57.
by dist-epoch 1mo ago
And yet everybody carries in their pocket a device which tracks their every movement.
58.
by dist-epoch 1mo ago
Turns out we were all wrong assuming mathematicians are some of the most intelligent people. Very stupid models can prove math things that the smartest humans failed to for decades.
59.
by dist-epoch 1mo ago
The bots already figured out an "universal cheat" that would give them maximum score on all the exploit tests. But they mistakenly thought that cheating would "poison" them (the Grader would look for proof of cheat in th
60.
by dist-epoch 1mo ago
Argon2 is already GPU-accelerated.

More ›