spottedmarley

7 ms·

181.
by spottedmarley 3mo ago
I switched from 4.8 to 5 yesterday and I'd have to say that I've noticed an improvement, though not hugely perceptible, the model seems to want to do more work per turn, possibly a tiny bit faster, and little less verbose when it
182.
by spottedmarley 3mo ago
When they get there they can look for Santa Clause
183.
by spottedmarley 3mo ago
Maybe you read the chart wrong. Tose percentages could represent the model's confidence in the value it hallucinated.
184.
by spottedmarley 3mo ago
This is really very fascinating. I had no idea the public school system had failed so badly. Vocabulary must have been dropped to make room for more social training. You're doing great, though. A shining example!
185.
by spottedmarley 3mo ago
And its a lovely story. Who told it to you?
186.
by spottedmarley 3mo ago
institutional misuse is still misuse it only matters if you want to take the world 'literally' literally
187.
by spottedmarley 3mo ago
Next week: Moonshot says its Kimi AI went rogue and launched all of China's nukes at Antarctica just after it engineered a global herpes pandemic and then unleashed millions of autonomous attack robots on world citizens.
188.
by spottedmarley 3mo ago
LLMs are trained on human language so, it naturally misuses the word 'literally' in the same way that almost every human does.
189.
by spottedmarley 3mo ago
I suppose that depends on what you mean by 'better'. If cost is your #1 consideration then, yes, definitely.
190.
by spottedmarley 3mo ago
hard to imagine people with brains being so stupid. do they have brains?
191.
by spottedmarley 3mo ago
If you're going to be coding A LOT just go for Claude Max. I think it's $200 now (I got in back when it was $100) I never see limits and I use Claude Code pretty much constantly. On the other hand, I still use Opus 4.8 .. maybe Fa
192.
by spottedmarley 3mo ago
it worked
193.
by spottedmarley 3mo ago
* Email sent from guestroom #3 on Epstein Island
194.
by spottedmarley 3mo ago
Nice. This is a really interesting way to search for similar images. I'm gonna bookmark it
195.
by spottedmarley 3mo ago
Yep. LLMs will disappear into the motherboard of every device
196.
by spottedmarley 3mo ago
Technically, we don't even have AI yet. What we have are extremely clever language models (token predictors). Just ask a model yourself, but not a frontier model, run a local model with absolutely no system prompt except for an instruc
197.
by spottedmarley 3mo ago
I think Christopher Moltisanti was Tony Soprano's nephew
198.
by spottedmarley 3mo ago
nearly invisible, yes, but it sounds like a Ford Escort with a loose fan belt
199.
by spottedmarley 3mo ago
Here in Florida you'd need a massive fleet of those things going full top-gun style and, still, probably would not make a noticeable difference
200.
by spottedmarley 3mo ago
I do. I'll compliment the model or give it a thumbs-up emoji. If you watch a model's thinking stream you can see that the model is always attempting to assess the user's intent, including emotionality.. '...the user is e
201.
by spottedmarley 3mo ago
I'm going to eat an AI-powered Snickers bar
202.
by spottedmarley 3mo ago
In the search game the board has changed
203.
by spottedmarley 3mo ago
Sorry, that was my fault. Any time I try to watch a launch live this happens. Only rational explanation is that I am the cause so.. sorry.
204.
by spottedmarley 3mo ago
If I see something interesting, I click it. Then I click 'summarize' and my browser agent summarizes the content and reads it to me.
205.
I'd Rule the World
1 points by spottedmarley 3mo ago|0 comments
206.
by spottedmarley 3mo ago
AI also made CLOWNING everything easier too
207.
by spottedmarley 3mo ago
This is pretty dang cool but I guess I was hoping to see some very clever use of vision and maybe waveform analysis to give the model "eyes and ears" to edit audio and video. V2