Why? Cuz Clinton Foundation.

And for ---- sake Galaxy Pete fire all those people and hire ones that aren't retarded.

This is trivial science to perform, put it on a pie set up the pie as your hotspot so it's its own isolated Network. Run the model. Ask it a math question x to the third plus y ^ 3 + z ^ 3 = 33 solve for all integers.

Enjoy the wait. Compare your results to the performance of API calls to deep sea more complex models, this new ---- from Alibaba, 01, grok, Etc

If you were as fortunate as me and only had to really work for 2 hours a day, you'd also have time for academic research, but I'm a unicorn.

If you actually had a computer science degree you would understand it is actually as good and fully tunable for performance unlike any of the current AIs which are more like a Runaway Diesel engine (google it) than a Ferrari.

#79 | Posted by GalaxiePete at 2025-01-29 05:06 PM | Reply | Flag:

I have more than 1 CS degree lol, do AI integrations, train models, etc. I've got DeepSeek on an airgapped PC and a raspberry pi. It's not that performant regardless of tuning. There's no substitute for racks of H100s and models with higher performance inferencing. It's cool if you want to edge deploy and don't mind how slow it is.

I would prefer to be wrong. Ukraine needs infantry, desperately.

Russia recruiting is surging, with lines around the block at centers. Everybody wants to get in on the contract cash grab before the war ends now that Trump is on office.

No, it's you.

Those are non-combat troops, and the talks are still ongoing. Ukraine needs combat troops, desperately.

Meanwhile, Trump continues to transfer Patriot missiles to Ukraine via Poland.

Wrong quote. SOrry, the dashboard(s) you were mentioning.

#72 | Posted by jpw at 2025-01-29 01:09 PM | Reply | Flag

Google it, there's a dozen and it's all the same public data. The only problem is the lag time between congressional trades and them filing the paperwork on it. In a few weeks we'll see who shorted Nvidia, and those will be the biggest crooks in Congress.

42.8 to 44.8, may bad. it's "only" 2000 points up.

She wasn't just selling stocks

#73 | Posted by hamburglar at 2025-01-29 01:35 PM | Reply | Flag:

She was buying other stocks. She even hedged on Nvidia after a larger previous position. She didn't unload completely, she knows it'll bounce back after the hysteria subsides.

If you really sold on Trump Fear, you're a panic seller and not represented by the DOW average which is up from 42.5 to 48.5 from inauguration day to now.

Nancy has only managed to be down 2% on her Nvidia holdings. Like the market doesn't matter.

There's many dashboards you can use to follow congressional stock trading.

The most active trader right now, by number of trades, is MTG.

It's public data.

Makes you miss the days of people with zero engineering education discussing Hyperloop like it wasn't a stupid idea.

If you think Ukraine has supply problems you haven't been keeping up.

They have recruiting problems. They did not do much to regenerate units and were creating new scratch battalions until recently, but it's not enough people. The order changes to send people to existing units to get experience. There's still not enough people.

Mike Walz has been signaling to Ukraine they have to lower their draft age.

Just waiting on the notification of her trade to rebuy on the dip. If you aren't, wtf are you doing?

If Nancy was an R instead of a D nobody here would be trying to defend a level of stock acumen that's only achievable by her and insider trading.

You don't need analysts yourself. Extra costs for lower returns. You just need to follow her trades on the plethora of sites that follow congressional stock action.

DeepSeek was released in September. It didn't just come out in the last 2 weeks.

Pelosi is one of the worlds greatest stock traders. It's been studied and is settled science no peer review has overturned. The only question is would she be that good without being in government? You're an idiot if you think so.

In the 90s we told Israel that Iron Dome would never work. Then it did.

A much more pressing problem, America doesn't make enough missiles to keep the Houthis suppressed. If an actual conflict against a near peer adversary breaks out, air defenses will be deleted in a few months or less.

So pithy. This is what it feels like to be a medical doctor watching people discuss taking horse dewormer.

And Nancy did not get out because the stock was overvalued. Nancy got out because her trades appear to be psychic. They appear to be psychic because she has massive contacts from being the chairman of the house intelligence committee. You don't need AI to trade like Nancy it's on a dozen different Congressional trade trackers, you just do what she does and you will flourish. She vastly outperforms the market average. The only question was why did she make that move and the answer came out two weeks later to us plebs.

Say it with me, test time scaling. Being able to use a novel TTS approach to dynamically adjust parameters during model training does not solve the basic inferencing problem that requires massive amounts of gpus to be fast and accurate. Deep seek is good at inferencing but it's not great, it's competing against platforms that are great.

To be fast and graded inferencing you have to have lots of Nvidia Cuda cores because they're optimized for linear algebra and clearly I can't be the only person in this forum that knows what matrix multiplication and row reduction is. Judging by this thread I might be.

The stock run is stock Movement by the same people who think Tesla is more valuable than all other car brands combined. Rationality is not on the menu.

If you have to look up what TTS is you're not ready for this conversation.

Man! Sounds like someone has a little sand ...

#22 | Posted by censored at 2025-01-27 05:04 PM | Reply | Flag:

When you see everybody going retarded over a TTS implementation, in GIANT CAPITAL LETTERS ON DRUGEREPORT, it's dumb.