Subquadratic – Introducing SubQ 1.1 Small

https://subq.ai/subq-1-1-small-technical-report

#HackerNews #Subquadratic #SubQ #Small #Update #TechNews #AI

SubQ is a new "subquadratic" LLM that can handle context windows of 12 million tokens. 12 million tokens is a massive amount of text, roughly equivalent to 9 million words or about 120 full-length novels. If this lives up to the claims, it's a game-changer. Wonder what the cost is of putting that many tokens in the context window, tho.

https://subq.ai/

#solidstatelife #ai #genai #llms #codingai #subquadratic

In our effort to put courses online, we continue lectures on Algorithmic Lower Bound Course. Now you can watch

Lesson 4-11: Algorithmic Lower Bounds by Mohammad Hajiaghayi - NP-Completeness and Beyond

(FEEL FREE TO SUBSCRIBE TO YOUTUBE @hajiaghayi FOR FUTURE LESSONS Premiering on WEDNESDAYS)

https://youtu.be/VZyffnAb1r0 (Lesson 4: 3-Partition Problem & Proving NP-Hardness)

https://youtu.be/4fCD9_1eQw0 (Lesson 5: Puzzle Problem NP-Hardness & 3-Partition)

https://youtu.be/FIyEj72-UJQ (Lesson 6: 3-SAT Problem & Proving NP-Hardness)

https://youtu.be/tbSJzaKx2pA (Lesson 7: Puzzle Problem NP-Hardness via 3-SAT)

https://youtu.be/voRVebBsh94 (Lesson 8: Fine-grained Subcubic Complexity: Part 1)

https://youtu.be/gRURSM6QARo (Lesson 9: Fine-grained Subcubic Complexity: Part 2)

https://youtu.be/qPw82bTAXkc (Lesson 10: Fine-grained Subquadratic Complexity 1)

https://youtu.be/C6j4avVkI7U (Lesson 11: Fine-grained Subquadratic Complexity 2)

#AlorithmicComplexity,

#3SAT,

#3Partition,

#subquadratic,

#subcubic,

#Finegrained,

#HardnessExploration,

#NP,

#PSPACE,

#NPComplete,

#LogSpace,

#ExponentialComplexity,

#ParallelComputation,

#PvsNP,

#NPSPACE,

#NonDeterministicSpace, hashtag

#SavitchTheorem,

#ComplexityClasses,

#Reductions,

#ImportantProblems,

#CommunicationComplexity, hashtag

#GeometricProblems,

#AlgorithmDesign,

#ComputationalComplexity,

#TheoreticalComputerScience,

#AlgorithmicLowerBounds

For comprehensive handwritten lecture notes on this course, visit the instructor's website:

http://www.cs.umd.edu/~hajiagha/
The course textbook "Computational Intractability: A Guide to Algorithmic Lower Bounds" by Demaine, Gasarch, and Hajiaghayi is available for free at:

https://hardness.mit.edu/