tcdent

15 ms·

211.
by tcdent 2y ago
Until the cookie banners go away there is zero ground to stand on.
212.
by tcdent 2y ago
adopting the mentality that AI agents are akin to russian spam bots is regressive mentality. your users will be interacting with your platform using partial automation in the very near future and if you think rate limiting or slowing their
213.
by tcdent 2y ago
Show me the Molex logo molded on that connector end and I'll believe you. It's all off-brand slop from companies that learned marketing by emulating western brands. Kids on the internet lap up circuitous threads about one brand be
214.
by tcdent 2y ago
This is the most informative assessment in this thread. You'd expect to see the capacity to be 125% as is common in other electrical systems. Ratings for connectors and conductors comes with a temperature spec as well, indicating the i
215.
by tcdent 2y ago
Yeah, how do people have these in their hands already? Everywhere I look is sold out.
216.
by tcdent 2y ago
Great reminder that there's more to adoption than just theory on paper; the practicalities, communities and a little bit of inexplicable magic are how new tech really takes off.
217.
by tcdent 2y ago
I've been paying almost $11 for them.
218.
by tcdent 2y ago
I like that I was able to style these with a super minimal theme in my menu bar that gives me a glance at what my system is up to and if it's outside normal parameters. Have been using this for about a year now without issue.
219.
by tcdent 2y ago
yeah, 'awful' is just the OP being dramatic, though I am looking forward to Astral's contribution
220.
by tcdent 2y ago
typed python is amazing for building understandable applications. I will never go back. you can obviously still use it untyped to hack together scripts, but I still find myself leaving annotations for clarity. I've been using it since
221.
by tcdent 2y ago
overlaying an additional conspiracy theory in skepticism of the current conspiracy theory
222.
by tcdent 2y ago
It's just bootstrapping. AGI will solve it.
223.
by tcdent 2y ago
Fab node size is not the only factor in performance. Physical limits were reached, and we're pulling back from the extremely small stuff for the time being. That is the evolutionary part. Revolutionary developments are: multi-layer waf
224.
by tcdent 2y ago
If you're expecting this device to stay relevant for 4 years you are not the target demographic. Compute is evolving way too rapidly to be setting-and-forgetting anything at the moment.
225.
by tcdent 2y ago
First indicator I've seen that Perl is not dead in years.
226.
by tcdent 2y ago
> In September 2023, the United States Marine Corps used an M72 LAW anti-tank rocket launcher fixed to a Go1 robotic dog during tactical training at the Marine Corps Air Ground Combat Center Twentynine Palms in California. https:/&
227.
by tcdent 2y ago
I used an 8500 as my personal machine for way too long. With a Sonnet G3 upgrade and maxed out RAM, it stayed viable for way longer than it should have.
228.
by tcdent 2y ago
Resource exhaustion can be super frustrating, and always feels like a third world slum situation when it happens. Like, why would operating systems allow themselves to run out of headroom entirely in this day and age?
229.
by tcdent 2y ago
A friend of mine ran a management company for a couple years, and indicated that with the talent they managed, it was almost never the actual person responding to DMs. The outsourced the interactions to third world "typists" that
230.
by tcdent 2y ago
If they were serious about AI they would have published TOPS stats at at least float32 and bfloat16. The lack of quantified stats on the marketing pages tells me Intel is way behind.
231.
by tcdent 2y ago
Do you have a roadmap for the future of the protocol? Is it versioned? ie. does this release constitute an immutable protocol for the time being?
232.
by tcdent 2y ago
The correlation with Iron makes me wonder if the polymerization of cooking oils is related. https://www.lodgecastiron.com/cleaning-and-care/cast-iron/sc...
233.
by tcdent 2y ago
These undergo additional fine tuning (QLoRA) using some or all of the original dataset, so they're able to get the weights to align to the nf4 dtype better, which increases the accuracy.
234.
by tcdent 2y ago
Not a doomer, but like, don't run this on your primary machine.
235.
by tcdent 2y ago
> Boeing was the main contractor blamed for SBINet's failure Why is this not surprising anymore?
236.
by tcdent 2y ago
Literally everything is trivial to jailbreak. The core concept is to pass information into the model using a cipher. One that is not too hard that it can't figure it out, but not too easy as to be detected. And yes, o1 was jailbroken s
237.
by tcdent 2y ago
Makes me think of my MacBook Air. The only part of the screen with fingerprints on it is over the front facing camera.
238.
by tcdent 2y ago
> from llama_models.llama3.reference_impl.model import Transformer This just imports the Llama reference implementation and patches the device FYI. There are more robust implementations out there.
239.
by tcdent 2y ago
Really hard to quantify the demand out there, since, thankfully, most of these people keep it out of my feed. But I have a feeling it's significantly more popular than we expect.
240.
by tcdent 2y ago
Depends on the architecture, but generally you just move through the layers linearly. Simple iteration. The number of layers, and the amount of time spent in each of them, makes me think any benefit from pre-loading the layer ahead is negli

More ›