Salesforce is overhauling its AI strategy, introducing generative AI
tools for tasks without human supervision and revising its pricing
model
Sign Up [1] |Advertise [2]|View Online [3]
TLDR
TOGETHER WITH [WorkOS] [4]
TLDR AI 2024-09-19
YOUR APP ENTERPRISE-READY (SPONSOR) [4]
You don't want to lose life-changing deals on a technicality.
Get WorkOS [4] and start selling to enterprises with just a few lines
of code:
Build your app with a complete User Management solution + SSO,
SCIM, and RBAC [4].
Integrate in minutes with modular and starightforward APIs.
Zero hassle across the board: UX your customers will love, quality
docs your developers will appreciate.
¨ User Management is free up to 1 million MAUs
Add WorkOS to your app in minutes ¨ [4]
¨
HEADLINES & LAUNCHES
SALESFORCE'S NEW AI STRATEGY ACKNOWLEDGES THAT AI WILL TAKE JOBS (5
MINUTE READ) [5]
Salesforce is overhauling its AI strategy, introducing generative AI
tools for tasks without human supervision and revising its pricing
model to $2 per AI-driven conversation. This shift aims to address
investor concerns about AI-induced job losses impacting
subscription-based revenues. The new AI tools offer higher efficiency
and autonomy compared to traditional copilots and chatbots.
QWEN 2.5 (12 MINUTE READ) [6]
An impressive array of open models that approach the frontier of
performance. Specifically, they have strong performance on code, math, structured output, and reasoning. The Qwen team has also released a
suite of sizes for a variety of use cases.
CREATE FULL WEB APPS WITH LLAMACODER (8 MINUTE READ) [7]
Together AI and Meta have partnered to make a tool that enables users
to build entire apps from just a prompt on the LlamaCoder platform. It
is akin to Claude Artifacts but purely created to show off the speed
of Together AI's inference engine.
¨
RESEARCH & INNOVATION
BITNET WITHOUT RETRAINING (18 MINUTE READ) [8]
The Hugging Face team, using a new technique called quantization warm
up, fine-tuned Llama 3 8B and maintained Llama 1 performance while
quantizing it to use 1.58 Bits per parameter.
1X WORLD MODEL (7 MINUTE READ) [9]
1x, a robotics company, has trained a video generation model that can
be used to simulate first person views of robotic activity. This can
be helpful for offline data generation and robot training.
DIAGRAM OF THOUGHT: ITERATIVE REASONING IN LANGUAGE MODELS (12 MINUTE
READ) [10]
The Diagram of Thought (DoT) framework introduces a new way for large
language models to reason by organizing ideas into a directed acyclic
graph (DAG). This method allows models to propose, critique, refine,
and verify ideas, improving logical consistency.
¨¨
ENGINEERING & RESOURCES
AI + PROCESS ORCHESTRATION = INTELLIGENT AUTOMATION (SPONSOR) [11]
While AI gets most of the spotlight, it needs to be thought of as a
part of the whole automation journey. Using AI alongside process
orchestration can unlock more efficiencies and increase the likelihood
of success with AI.
Learn how to get started: Read _The Ultimate Guide to Intelligent
Automation_ [11] by Camunda
ZML INFERENCE (GITHUB REPO) [12]
ZML is a lightning fast inference engine written in Zig that is
designed with performance in mind. It can be used for a variety of
models, although it requires some customization to get working for a
new architecture.
ADVERSARIAL ATTACKS ON NAVIGATION AGENTS (GITHUB REPO) [13]
This repository introduces a new attack method for embodied
navigation agents that attaches transparent patches with learnable
textures to target objects.
GRAPH ANOMALY DETECTION: A SURVEY (29 MINUTE READ) [14]
This paper offers an in-depth review of deep learning methods,
particularly graph neural networks (GNNs), for detecting anomalies in
graph data. Researchers introduce a novel taxonomy of approaches,
analyzing GNN designs, proxy tasks, and anomaly measures.
¨
MISCELLANEOUS
SOCIALAI OFFERS A TWITTER-LIKE DIARY WHERE AI BOTS RESPOND TO YOUR
POSTS (9 MINUTE READ) [15]
SocialAI, a new iOS app, offers a social media experience solely
populated by AI-powered bots, eliminating human interaction. Users can
share thoughts and receive endless, tailored AI-generated responses,
choosing from categories like "supporters" or "critics." Founder
Michael Sayman aims to provide a private, interactive space leveraging
large language models for diverse feedback.
ACADEMIC TEXT PARSING (GITHUB REPO) [16]
AceParse is a dataset designed to improve the parsing of structured
texts in academic papers, including formulas, tables, and complex
sentences.
JUMPING OVER AI'S UNCANNY VALLEY (9 MINUTE READ) [17]
This article explores the Uncanny Valley theory, which suggests that discomfort with almost-human AI could hinder its adoption. It examines
recent advancements in AI that bring this psychological phenomenon
into focus, questioning its potential impact on AI's future. The piece concludes by suggesting that AI's best use may be in a complementary
role rather than as a human replacement.
¨
QUICK LINKS
THE DATA THAT MAKES REALISTIC AI VOICE GENERATION POSSIBLE (SPONSOR)
[18]
Learn how human-curated language data supports innovative
text-to-speech advancements and gives content creators more control
over voiceover. Watch the recorded webinar by Oxford Languages and
WellSaid [18]
MERCOR'S $30M SERIES A (1 MINUTE READ) [19]
Mercor raised $30M from Benchmark to build an AI recruiter.
CANCER DETECTION (GITHUB REPO) [20]
SkinMamba is a hybrid model that combines CNN and Transformer
techniques to improve skin lesion segmentation for early cancer
detection.
AMAZON ALEXA CAN NOW BE CONTROLLED BY THOUGHT ALONE - THANKS TO THIS
BRAIN IMPLANT (2 MINUTE READ) [21]
Synchron has enabled an ALS patient to control Amazon's Alexa with a
brain implant, bypassing the need for voice or touch.
Love TLDR? Tell your friends and get rewards!
Share your referral link below with friends to get free TLDR swag!
https://refer.tldr.tech/21532aea/2 [22]
Track your referrals here. [23]
Want to advertise in TLDR? ¨
If your company is interested in reaching an audience of AI
professionals and decision makers, you may want to ADVERTISE WITH US
[24].
If you have any comments or feedback, just respond to this email!
Thanks for reading,
Andrew Tan & Andrew Carr
If you don't want to receive future editions of TLDR AI, please
unsubscribe from TLDR AI [25] or manage all of your TLDR newsletter subscriptions [26].
Links:
------
[1]
https://tldr.tech/ai?utm_source=tldrai
[2]
https://advertise.tldr.tech/?utm_source=tldrai&utm_medium=newsletter&utm_campaign=advertisetopnav
[3]
https://a.tldrnewsletter.com/web-version?ep=1&lc=df5c4ca8-734c-11ef-b5ad-9577e7a7de79&p=2d75fb62-7674-11ef-88da-0fdf13db5557&pt=campaign&t=1726752221&s=2bd34adceda6d3b0b59575b2231858200420960d4f7c30f48d961878cfe79f08
[4]
https://workos.com/?utm_source=tldr&utm_medium=newsletter&utm_campaign=q22024
[5]
https://links.tldrnewsletter.com/cCSwJ2
[6]
https://qwenlm.github.io/blog/qwen2.5/?utm_source=tldrai
[7]
https://links.tldrnewsletter.com/6Z4s1A
[8]
https://huggingface.co/blog/1_58_llm_extreme_quantization?utm_source=tldrai [9]
https://www.1x.tech/discover/1x-world-model?utm_source=tldrai
[10]
https://arxiv.org/abs/2409.10038v1?utm_source=tldrai
[11]
https://page.camunda.com/wp-intelligent-automation-guide?utm_medium=paid_leadgen&utm_source=tldr&utm_campaign=Guide.UltimateGuidetoIntelligentAutomation.24Q3.EN
[12]
https://github.com/zml/zml?utm_source=tldrai
[13]
https://github.com/chen37058/physical-attacks-in-embodied-navigation?utm_source=tldrai
[14]
https://arxiv.org/abs/2409.09957v1?utm_source=tldrai
[15]
https://techcrunch.com/2024/09/17/socialai-offers-a-twitter-like-diary-where-ai-bots-respond-to-your-posts/?utm_source=tldrai
[16]
https://github.com/JHW5981/AceParse?utm_source=tldrai
[17]
https://every.to/p/jumping-over-ai-s-uncanny-valley-4b1c3436-b424-4a62-b563-50e1469bba6c?utm_source=tldrai
[18]
https://languages.oup.com/products/webinar-what-can-curated-pronunciation-data-do-for-text-to-speech/?utm_source=tldr+ai&utm_medium=newsletter&utm_id=ws+on+demand+webinar+19092024
[19]
https://threadreaderapp.com/thread/1836435248592376149.html?utm_source=tldrai
[20]
https://github.com/zs1314/skinmamba?utm_source=tldrai
[21]
https://www.zdnet.com/article/amazon-alexa-can-now-be-controlled-by-thought-alone-thanks-to-this-brain-implant?utm_source=tldrai
[22]
https://refer.tldr.tech/21532aea/2
[23]
https://hub.sparklp.co/sub_a89cbcf98f89/2
[24]
https://advertise.tldr.tech/?utm_source=tldrai&utm_medium=newsletter&utm_campaign=advertisecta
[25]
https://a.tldrnewsletter.com/unsubscribe?ep=1&l=eedf6b14-3de3-11ed-9a32-0241b9615763&lc=df5c4ca8-734c-11ef-b5ad-9577e7a7de79&p=2d75fb62-7674-11ef-88da-0fdf13db5557&pt=campaign&pv=4&spa=1726750844&t=1726752221&s=c27c8167c74e0a265ae92299731365c6cd988f90fa5c9d818bc2d3ee5a2d9b0d
[26]
https://tldr.tech/ai/manage?email=tldr%40synchro.net
---
þ Synchronet þ Vertrauen þ Home of Synchronet þ [vert/cvs/bbs].synchro.net