Anders G Frøseth

@DavidDeutschOxf Well, «philosophers» like Curtis Yarvin who have a strong influence on important actors behind Trump wants to reorganize the US based on ideas that even predates the renaissance. So they are well under way to destroy enlightment institutions

Doing the right thing on this issue and its broader manifestations is more important than saving Ukraine, even though the suffering is far greater in the latter. 😢 😢 😢 😢 😢 😢 Because it’s more important that the Enlightenment not be destroyed in the US than Ukraine.

@togelius Agreed! And his tweet hints at obscuritanism (at best)

I've never understood the appeal of Yudkowsky, nor the perceived validity of his reasoning, probably because I haven't taken the time to self-indoctrinate into lesswrong rightthink. I would happily read a paper (or a few) that explained his key positions and arguments in language that I could understand. But, alas, he considers it a waste of time to write such a thing, and nobody else is capable of doing it to satisfaction. Ah well.

@nathanbenaich @airstreetpress Thanks and thanks for sharing! In my experience, in early deep tech investing, the idea of a funnel and how many deals that are incoming is not really interesting. As this is the most private and competence driven market that exists in any sense of finance theory 🙂

New on @airstreetpress - last year we evaluated ~450 opportunities and countless even earlier stage ideas In the end, we made 3 seed investments The biggest single reason for passing was that ideas were unexciting So, what makes for an exciting opportunity? Thread!

@togelius Cars in the 1830s

I love technology and am generally an optimist, but clearly not all tech is destined to shape our future. What are examples of tech that was once thought to conquer the world, but then fizzled? Particularly if it's not because of competition, but the world just wasn't interested.

@olexandr The real AI bump comes when AI can be used to predict clinical trail outcomes

Quite depressing trend. When do we see the AI bump?

@AmirBaratiF Congrats!

Feeling grateful, two papers accepted at #neurips2023. Our Scalable Transformer paper on PDE surrogate modeling (arxiv.org/abs/2305.17560) and Our Symbolic Regression paper with Transformer planning for math (arxiv.org/abs/2303.06833). Congratulations to my PhD students.

@vijaypande Well. That is some coincidence :)

The cookie told me this:

@leecronin @ChemifyX @TriatomicCap Congrats Lee! Well done!

So excited that @ChemifyX today announced $43 million of funding including a series A led by @TriatomicCap to ensure that chemistry & chemical discovery becomes digital, building a century defining technology. #chemputation

Great paper alert! @A_Aspuru_Guzik @mmbronstein and many more arxiv.org/abs/2307.08423

@realgeordierose Yes, by continuing the great work at @TheSanctuaryAI ! 🙂

Finse Hotel guests @HotelFinse1222 @MarkHamill

@AmirBaratiF @YuyangW95 @changwenxu Very nice! Congrats :)

Check out our #Transpolymer paper in Nature Computational Material on using #GPT and #Transformers and #Pretraining for polymer property prediction. SOTA on all polymer datasets and tasks across the board: nature.com/articles/s4152… @YuyangW95 @ChangwenXu #GPT4 #ChatGPT #GPT4Science

@curiouswavefn Sounds like a good recipe for dying young, including all sorts of physical and mental health problems along the way. Not much to celebrate :)

And end the day with an Oppenheimer Martini: A potent 4 oz of gin and a smidgeon of vermouth in a chilled glass whose rims are dipped in honey and lemon juice. The father of the atomic bomb liked his martinis dry and potent and his guests drunk and happy.

Since it's Oppenheimer's birthday, let's celebrate it by eating "black and tan", a high-energy breakfast concoction whipped up by him when he was a student at Harvard that would keep him going all day long. Toast + thick layer of peanut butter + thick layer chocolate sauce.

@vijaypande @sundarpichai @60Minutes Just a thought. One of the major discrepancies with the current models wrt the brain might that connections between nourons are engineered or static. And then there are the issues of activation (does spiking make a difference) and the separation of memory from compute.

And @sundarpichai is spot on when he mentions to @60Minutes that we don't understand the human brain. I'd add that *people* are the true blackbox, whereas AI can deeply probed and its blackbox made transparent. /4 nytimes.com/2018/01/25/opi…

AI is an emergent property from neural networks. But that doesn't mean we don't understand it. We know how the algorithms work. A greater issue that emergent properties are new to many computer scientists and just weird. But not new to scientists. /1 twitter.com/60Minutes/stat…

@nick_farina Well, yes. Not only speaking but walk the talk when it comes to conferences. Othetwise, this is what you do as a founder, @nick_farina’s. Find your market! Also happy to do a DM if interested. 🙂

those of us in QC leadership roles need to spend less time speaking at QC conferences and more time at general conferences telling the rest of the world what the hell we are doing (and why it matters)

@togelius The one which is an explicit and devastating critique of Kant. Would also be happy to receive the modern version.

@andersagf The Artificial General Intelligence one? I'm happy to sign it 😀

The book @togelius. Really looking forward to it. Will you send out signed copies to your fellow ignorant Vikings?

@vijaypande 1) And one could get more specific pretraining 2) This is also major moat, for the companies who can get it right

@andersagf yes and no. 1) you can do a lot more with less, few shot methods are everywhere (and LLMs are inherently few shot in nature at scale) 2) successful companies have GTM built around access to data, ideally access to data unique to the company

I don't care about all the details in mall retrofits for bio, just as long as there's a strong core facility and an Orange Julius

@nick_farina Well. I hope you are on fire, and that you will keep that fire burning through the hype :-)

And yeah, given history, it’s kinda weird how the Chicago tech community uses the brand “fire” for everything but YOLO.

Chicago is my hometown, so always great to see EeroQ recognized as a local leader, this time on the @bizjournals “Fire 50” list! bizjournals.com/chicago/inno/s…