Chris @cdossman Twitter profile | Pikagi

Pikagi

Chris

@cdossman

1,479

Followers

1,343

Following

452

Media

4,681

Statuses

Co-founder @ | Building Artificial Minds | Google Developer Expert in ML

https://t.co/QQqwa8FSMv

Joined December 2009

Don't wanna be here? Send us removal request.

Pinned Tweet

@cdossman

Chris

7 months

Self-Discover Prompting Implementation: Currently seeing a 86.6% on a random subset of BBH, which puts me in line with the results seen by Deep Mind. @lateinteraction I tried out many of the functions built in to DSPy. If you think it makes sense I can

Tweet card media

Self_Discover_Implementation_Using_DSPy

Colaboratory notebook

colab.research.google.com

6

28

103

Last Seen Profiles

@kubotaice

@rolixdragon

@bokeplokalmalam

@ggukvodoo

@CoachKait_

@GuzKhanOfficial

@RAJKUMARBHAGA16

@Testerman_42

@dwjr_0

@bokeplokalmalam

@StwDood38499

@MMPowerFdn

@PHILDEP

@ElodieHuchard

@eduardoanv

@PKukanja

@michaelsb55

@huanghuangxuan1

@CaptCalcium

@xexr00

@gajyago

@dylanrefuge

@rubytogether

@jacques_chester

@HarvardFootball

@toddtulgestke

@waifumika_

@SSN_Griz

@kevinthai89

@sussooss

@Jaz9_

@Abti00

@jandakembangstw

@devindadude_4

@lutherking

@DPN_PETANI

@cdossman

Chris

10 months

@ylecun I think you might be discounting billions of years of evolution that might have imparted some type of "hardware" pre-training for our brains.

15

9

426

@cdossman

Chris

11 months

@antoniogm You being in Portugal is bringing high price dystopia to the Portuguese people 1 remote worker at a time.

12

4

392

@cdossman

Chris

2 years

@dkbrereton @fchollet This image is becoming more and more true every passing day.

Tweet media one

2

26

286

@cdossman

Chris

2 years

An early demo of our action engine. LLMs are the on-ramp to solving real-world problems. Soon we will be able to have always on a personal assistant that can help solve a wide range of daily tasks. "Find me the cheapest floor seat tickets to Madonna show in Seattle."

6

20

262

@cdossman

Chris

2 years

@hlntnr @anthrupad My personal favorite version of this meme.

Tweet media one

3

10

174

@cdossman

Chris

11 months

@antoniogm Which are increasing rapidly as the influx of high paid remote workers flood the cities and country side. Your highly paid because you ideally live in a high cost of living area. Those who move to low cost living areas drive prices up for locals who have no chance to earn at the

12

0

168

@cdossman

Chris

6 years

So glad I had my cheat sheets today. Makes my job a lot easier

Tweet card media

Top 6 Cheat Sheets Novice Machine Learning Engineers Need

Earlier this week a saw a post of some excellent cheat sheets that I wish I had access to when I first started learning ML. It was nice to…

link.medium.com

3

44

139

@cdossman

Chris

10 months

Spent the weekend trying to build a version of @dicermarketing with GPTs. Learned a ton and I'm sharing it below 👇including how to turn any GPT in to a upsell machine. 🚀 Introducing MetaGPT: Your Ultimate AI Co-Pilot for Meta (Facebook) Ads! 🤖

Tweet card media

ChatGPT - MetaGPT : Meta Ads AI Marketing Co-Pilot

Expert in Meta advertising that can improve your ROI. Official Meta GPT built by dicer.ai

7

10

103

@cdossman

Chris

2 years

Launched an LLM-powered question-answering service that is very accurate through the use of prompt chaining and access to a human-verified database of known facts. Thanks, @weaviate_io , for the support and your fantastic vector database!

4

20

80

@cdossman

Chris

4 years

BERTweet: A pre-trained language model for English Tweets Receive #AI updates for FREE by subscribing to our newsetter Research link:

Tweet media one

4

12

51

@cdossman

Chris

2 years

Shout out to the fantastic companies, products, and people that help make it possible to build these awesome tools. @weaviate_io Vector Database @LangChainAI Build LLM products @OpenAI LLM API OG @SeleniumHQ Browser API @mathemagic1an Papers and news

2

7

49

@cdossman

Chris

2 years

Beginners: 3 LLM projects for beginners: 1. Search via a vector database like @weaviate_io 2. Create a prompt to generate tweets from a set of examples using @OpenAI 3. Fine-tune a less expensive model to perform the same task in project 2 using @OpenAI

2

5

44

@cdossman

Chris

2 years

We are looking for a UI/UX designer who wants to work on a high-growth company focused on creator monetization tools. If interested and want to join, DM me.

26

0

24

@cdossman

Chris

2 years

@JoshwaBail @dkbrereton @fchollet Found it! @anthrupad was the original artist

1

0

30

@cdossman

Chris

2 years

Been using for a while and it really helps us build our LLM applications. Their one click fine tuning pipeline is great and I can’t wait for support for Open-source models

@RazRazcle

Raza Habib

2 years

Is LLM finetuning worth it? If you know what you're doing, finetuned models can be 30x smaller❗️without losing performance. This can unlock applications that would otherwise be too expensive or slow. STaR is a great example of doing it right. (DIY instructions at the end)

Tweet media one

10

125

877

1

9

28

@cdossman

Chris

2 years

@heykahn You forgot which lets you speak to your data, join data across multiple sources, and create visualizations. It’s in open alpha and completely free to use

Tweet media one

Tweet media two

2

3

25

@cdossman

Chris

5 years

Meta-Dataset: a new benchmark for training and evaluating models that is large-scale, consists of diverse datasets, and presents more realistic tasks by @Eleni30fillou @TYLERZHU3 @imkelvinxu @kswersk @hugo_larochelle

1

12

25

@cdossman

Chris

1 year

Excited for “The Future of LLMs” Thursday July 6th: @bobvanluijt (CEO @weaviate_io ) @hwchase17 (CEO @LangChainAI ) Angela McNeal (Ex-Palantir AI lead) @johnpdickerson & @apwenchel ( @itsArthurAI ) @jefranklel (CSO @MosaicML ) Moderated by @alexrkonrad

Tweet media one

0

3

25

@cdossman

Chris

2 years

@hlntnr @anthrupad Artist: Anna Husfeldt, CC-BY SA 3.0

0

0

20

@cdossman

Chris

2 years

"Great outputs from on FB performance marketing data" -- @elliottaj Jenee AI helps people converse with their databases, files, Google Analytics, and more. Collect and Join Data Across Multiple Sources.

Tweet media one

0

1

19

@cdossman

Chris

5 years

Object Discovery with a Copy-Pasting GAN Amazing research work by @relja_work and @AndrewZisserman

Tweet card media

AI Scholar: GANs for Object Discovery

Generative adversarial networks (GANs) are tremendously being used for image generation from scratch including image translation…

0

7

16

@cdossman

Chris

2 years

@iamthatlucian @danlovesproofs Yes it’s something that many startups have figured out and they have used it to generate fantastic results. I’ve personally been involved with a team building a factual ChatGPT. You can use the basic version here at

2

0

19

@cdossman

Chris

2 years

Finished Squad 2.0 testing for the Maryanne question-answering system. 98.1% accuracy on the evaluation set (8% better than the nearest competitor). LLMs have a capability overhang described by @jackclarkSF here are some advanced Usage patterns for LLMs Thread Below👇

2

7

18

@cdossman

Chris

11 months

Great panel at signal vs. voice event hosted by @mprkhrst Exploring the impact of generative AI in art and the world. @elevenlabsio @SabrinaHalper

Tweet media one

1

2

17

@cdossman

Chris

2 years

@LangChainAI meet up in NYC. Great crowd and energy!

Tweet media one

0

4

17

@cdossman

Chris

2 years

NYC founder dinner for those building LLM applications! Thanks for hosting @doppenhe @chamii22

Tweet media one

1

0

16

@cdossman

Chris

2 years

Introducing our latest product: an AI data scientist that never shies away from a challenge! Our team created an incredible algorithm that recognizes when it's missing key data and seeks user input to solve its blockers. Proud to have such a brilliant

Tweet media one

1

5

15

@cdossman

Chris

2 years

@alexgraveley Met up with a brilliant researcher @LoganGWright1 working out of @Cornell University. My best guess for getting LLMs on the edge would be his physics-aware neural networks. Amazing novel work using non-linearities in physical systems to do ML work

Tweet card media

Deep physical neural networks trained with backpropagation

Nature - A hybrid algorithm that applies backpropagation is used to train layers of controllable physical systems to carry out calculations like deep neural networks, but accounting for real-world...

1

1

16

@cdossman

Chris

5 years

Researchers share their experiences with computational text analysis. How we do things with words: Analyzing text as social and cultural data. @SimonDeDeo @jacobeisenstein @dmimno @jfwinters @xrysoflhs @RebekahKTromble

Tweet card media

AI Scholar: Computational Text Analysis — Analyzing Text as Social and Cultural Data

Providing Best Practices for Working with Thick Social and Cultural Concepts

0

5

15

@cdossman

Chris

2 years

10x cheaper and better performance. Seems like I need to update my timeline every week. Leave some innovation for the rest of us @sama 😂

Tweet media one

2

1

14

@cdossman

Chris

5 years

New Variational Bayesian Context-Aware Representation (VBCAR) model for grocery recommendation that outperforms existing state-of-the-art methods by @mengzaiqiao @richardm_ @craig_macdonald @iadh

Tweet card media

GitHub - mengzaiqiao/VBCAR: Variational Bayesian Context-aware Representation for Grocery Recomme...

Variational Bayesian Context-aware Representation for Grocery Recommendation - mengzaiqiao/VBCAR

0

5

15

@cdossman

Chris

1 year

@t3dotgg Bought in to Google music and rode the wave. $8/m for YouTube red. It’s awesome. Feel like thee people who bought a home when you could for 30k

Tweet media one

0

0

13

@cdossman

Chris

11 months

@saigonandy1 @antoniogm Rising prices and a lack of economic activity that was promised with an influx of tax free remote workers 🫠

3

1

14

@cdossman

Chris

2 years

The home-runs don’t stop coming. Great work on the site. Needed to use Flippa recently for a transaction and it’s like using a site from the early 2000’s. Any plans to broaden the deal types on microaquire? Partial transactions are a huge and great for biz dev

@agazdecki

Andrew Gazdecki

2 years

Most startups don't have a clean profit and loss statement. Most buyers looking to @microacquire require a clean profit and loss statement. So, we're launching a P&L builder to helps founder show the true financial health of their startup in real time. Stay tuned, tres commas.

40

36

433

2

3

15

@cdossman

Chris

1 year

Launching with a new agent feature. Control every aspect of the Dicer application through chat. - talk to your marketing data - plan new campaigns and creatives - launch better ads

3

0

12

@cdossman

Chris

2 years

Thank you @microacquire for the great swag. As if successfully selling my company wasn’t awesome enough, you guys coming through with this nice gift. @agazdecki is building an amazing platform 🙌

Tweet media one

1

2

11

@cdossman

Chris

2 years

@nearcyan it’s still a engineering demo but we are improving it regularly. MVP coming in the end of February.

1

1

14

@cdossman

Chris

2 years

@MetaverseJosh @Scobleizer @panxueni

Tweet media one

0

0

11

@cdossman

Chris

1 year

NYC AI scene is alive and well! Looking forward to this all star panel.

Tweet media one

Tweet media two

0

1

13

@cdossman

Chris

2 years

Got the chance to meet up with @Jonpon101 . Thanks for hosting me and sharing your story and product! If your building with LLMs and haven’t tried @promptlayer try it today. Maintain a log of your prompts and OpenAI API requests. Track, debug, and replay old completions.

Tweet media one

1

2

12

@cdossman

Chris

2 years

Another day of me and the boiz building LLM applications.

Tweet media one

2

0

12

@cdossman

Chris

2 years

Just passed 100k answered questions and 1k app downloads. Markdown formatting launch last week, this week we just tested LLM action engine, next week we will introduce direct user feedback system. LLMs are the on-ramp to solving real world problems. Imitation learning at scale.

Tweet media one

Tweet media two

Tweet media three

Tweet media four

0

5

12

@cdossman

Chris

11 months

@RileyGiuliano @antoniogm Perhaps for the young and able, what of the elderly or those on fixed incomes. Many stories of people going homeless or being forced to move from their home towns because it has become a tik tok hotspot

2

1

11

@cdossman

Chris

2 years

Bring on the collaborative foundation models

@DynamicWebPaige

👩‍💻 Paige Bailey

@DynamicWebPaige

2 years

"Deep learning requires increasing compute to train SOTA models, which is limited to few large companies. To alleviate this disparity, smaller groups may pool their computational resources and run collaborative experiments that benefit all participants."

Tweet media one

4

3

18

1

2

12

@cdossman

Chris

1 year

Tweet media one

0

0

8

@cdossman

Chris

3 years

Great thread of multiples of different businesses

@agazdecki

Andrew Gazdecki

3 years

For years, you’ve been told by brokers and intermediaries what multiples are fair for startup acquisitions. Today, we’d like to share with you @microacquire 's 2021 analysis of acquisition multiples in the first of our biannual multiple reports. Details below. Let's dig in.

Tweet media one

32

103

661

1

5

11

@cdossman

Chris

2 years

Modern AI Developer Toolbox 🔧 💽 Datasets + Model Directory @huggingface 📝 Vector Databases @weaviate_io 🖌 LLM API @OpenAI 💻 Demo your creation @Gradio 🚀 Market and Launch your new project on Twitter

0

2

11

@cdossman

Chris

2 years

@ash_lmb Work at a few different companies before you start your own. Too many young entrepreneurs run before they walk.

1

0

10

@cdossman

Chris

1 year

@weaviate_io to the rescue

@Exploringfornow

Yasir

@Exploringfornow

1 year

Just paid $1,000 for a month on @pinecone with no real usage. I'll give Pinecone benefit of the doubt to say it's a poorly designed pricing page. Nonetheless beware of predatory pricing strategies, esp if you're new to building generative AI products. Pinecone's pricing page: 1)

Tweet media one

55

28

402

1

0

10

@cdossman

Chris

2 years

@heykahn It's not just a SQL writer. The system has been trained to write code to interact with all types of data stores. CSV files, MySQL databases, and google analytics (coming soon). In the next few weeks, we will expand to include much more.

Tweet media one

1

0

9

@cdossman

Chris

1 year

Excited to share the latest on FAQx's goal-oriented action agent! Our agent is a true self-starter, able to learn and self-correct without any prior website fine-tuning. The current version only controls the browser, but the next iteration will be able to control your computer

2

5

10

@cdossman

Chris

3 years

Bringing the power of liquidity to Africa! Way to go @microacquire

@agazdecki

Andrew Gazdecki

3 years

We've gone global at @MicroAcquire ! Excited to announce we've partnered with @pariti_io , a community serving entrepreneurs in over 40 African countries. Together, we aim to provide the infrastructure and support that helps startups in underserved markets get acquired. 1/

Tweet media one

28

69

370

0

2

10

@cdossman

Chris

5 years

Build your understanding of DQN in three easy steps #reinforcementLearning #deeplearning

0

7

7

@cdossman

Chris

3 years

Yes! Finally, someone is helping make selling companies as easy as selling products on an e-commerce site.

@agazdecki

Andrew Gazdecki

3 years

@microacquire Here's a higher resolution demo of the @microacquire "LOI builder". Our goal is to streamline and simplify the creation of expensive legal docs for acquisitions. Feedback welcomed!

5

2

23

1

3

10

@cdossman

Chris

11 months

Holy cow! #NYCtechweek was exhausting after a week of events, but the event hosted by @SabrinaHalper at @HOFCapital was the cherry on the Sunday. Shout out to the great entrepreneurs and investors who shows up that made it amazing. @riomensah @AJ_Cats_ @yasmingao_

Tweet media one

Tweet media two

1

0

10

@cdossman

Chris

1 year

VC investing process: Standford? Deep Mind? / \ yes no / \ does founder meh speak v fast? / \ yes no / \ lead small check

0

1

9

@cdossman

Chris

2 years

Great time at the @Gradio / @huggingface hackathon in NYC. Thanks for the amazing time

Tweet media one

Tweet media two

0

2

8

@cdossman

Chris

1 year

📣 We're excited to announce the launch of our revamped Dicer(.)ai website! We'd love to hear your thoughts. What do you like most? Any suggestions for improvement? 🙏🌟

Tweet media one

1

1

9

@cdossman

Chris

2 years

Dashboard and the new graphing library are set up. 1.5 weeks until MVP is launched, try out the engineering demo at Talk to your databases, Google analytics, CSV files and many more to come.

Tweet media one

Tweet media two

3

1

8

@cdossman

Chris

1 year

First time I've been this excited in a while

Tweet media one

0

1

9

@cdossman

Chris

10 months

@ylecun If not directly in stored data, then perhaps in the form of a training recipe over fit to our environment. Developmental milestones are important for children. Discreet modes of operation that our bodies systems (Brian’s, bodies, hormones, microbiom, etc) take in order to

3

0

8

@cdossman

Chris

1 year

Great time at the GPT/LLM hackathon in NYC this weekend. Our team built automated networking system for the hackathon. Cloned people with LLMs and scraped internet data. Simulated conversations between all the attendees. Works decently well!

1

0

8

@cdossman

Chris

2 years

🔥New Feature Update 🔥 ✅ Multivariate Graphing ✅ Natural Language Feedback ✅ Sassy mode 😈 Jenee AI is trained to connect to your database, write code for data wrangling, and deliver visual insights to help you answer tough questions. No Tech

Tweet media one

Tweet media two

1

1

8

@cdossman

Chris

2 years

Got to meet @MrHunterBrooks cofounder of . Amazing founder and great product. It’s an innovative prompt engineering workspace for developers. Great UI for building on top of @LangChainAI and tons of other useful features. Try it this weekend.

Tweet media one

0

4

8

@cdossman

Chris

2 years

End of another successful AI workshop weekend. The theme? Enterprise AI Agents and a new prompt engineering techniques to get the most of 32k tokens. Next 2 months going to be wild. @elliottaj @marcobianco @made_by_connor @povgoggles need to get you up for the next one

Tweet media one

Tweet media two

0

2

7

@cdossman

Chris

2 years

@nntaleb Really love your books and I’m sorry that ChatGPT failed you. Bing GPT is a bit better because it is connected to the internet, but it often hallucinates as well. Checkout for a factualGPT experience. It is connected to the internet and a database of

0

2

8

@cdossman

Chris

2 years

Kind of lame how Meta is getting so much attention from a paper that describes what people on twitter have been doing for 6+ months

@_akhaliq

AK

2 years

Toolformer: Language Models Can Teach Themselves to Use Tools introduce Toolformer, a model trained to decide which APIs to call, when to call them, what arguments to pass, and how to best incorporate the results into future token prediction abs:

Tweet media one

32

276

1K

1

0

6

@cdossman

Chris

3 years

This guy is so savage. He is breaking apart what media has force fed young entrepreneurs all their life. Keep it up and save the future from the mistakes of the past.

@agazdecki

Andrew Gazdecki

3 years

One of my life goals is to build a startup the Silicon Valley way. 1. Get into YC 2. Have cofounders 3. Raise a Series-D 4. Add “IPO” to company values 5. Burn $5m+ monthly 6. Own 10% of my company 7. Say “double click” a lot 8. Hire expensive PR agency 9. Billboard on the 101

47

41

686

1

0

7

@cdossman

Chris

2 years

@passionfingerz @danlovesproofs Exactly you build corpuses of examples, information, or other things and make them searchable for use in prompts to improve results.

0

0

7

@cdossman

Chris

2 years

Tweet media one

Tweet media two

2

2

6

@cdossman

Chris

3 years

Getting ready for my day with a comfy boost from @microacquire . Thanks for the socks!

Tweet media one

0

1

7

@cdossman

Chris

5 years

Helping robots learn from large-scale offline task demonstrations through IRIS by @AjayMandlekar @drfeifei @animesh_garg Receive #AI research updates for free 👉 #ai #reinforcementlearning #educateai

Tweet media one

0

1

6

@cdossman

Chris

3 years

Selling a company is one of the hardest things emotionally and mentally I’ve ever had to do. So many ways to screw up and very few paths to success. Even with platforms like @microacquire or @Flippa it still requires a ton of patience, deal experience, and trust.

2

1

7

@cdossman

Chris

2 years

Had a great conversation with @bobvanluijt and @marcobianco . Hour-plus conversation talking about: Pairing vector databases and LLM to solve the data hallucination problem How vector databases are powering the future of intelligent systems And much more. Can't wait to release.

2

3

7

@cdossman

Chris

2 years

This Uber driver is taking his pitch game to the next level 🚗💸 Who needs Shark Tank when you have a captive audience in the backseat?

Tweet media one

1

1

7

@cdossman

Chris

2 years

Building products with LLMs? Make sure that making training data is as simple as using a product and giving feedback. Every user interaction can provide valuable fine-tuning data, making every future interaction higher quality. It's human feedback loops all the way down.

0

2

7

@cdossman

Chris

7 years

@jGage718 Wrote an amazing article on the number 1 issue with building #deeplearning #AI powered companies. Deployment is a difficult problem to solve. Good space for companies to serve with Ops software/services.

0

3

6

@cdossman

Chris

2 years

@ptsi @RazRazcle is already connected to wolfram alpha and handles math very well. Just passed 100k questions answered after 1 month.

0

1

6

@cdossman

Chris

11 months

Rise and shine, the future is waiting to be built

Tweet media one

0

1

6

@cdossman

Chris

7 years

What happens when chat bot meets Style transfer? Beautiful and sometimes creepy crowd sourced art. #deeplearning #chatbots #wechat

Tweet media one

Tweet media two

Tweet media three

Tweet media four

0

3

6

@cdossman

Chris

2 years

We are gearing up to fully launch in the next week or so. Talking to your data, joining across sources, Automated data science workflows, and a ton of connections coming in the next 2 weeks.

Tweet media one

1

0

6

@cdossman

Chris

8 months

@francoisfleuret It’s happening in industry and in labs already. We built a huge dataset of high quality video descriptions for marketing specifically using this and other techniques.

1

0

4

@cdossman

Chris

5 years

I just published Object Detection Accuracy (mAP) Cheat Sheet @kdnuggets #MachineLearning #AI

0

0

5

@cdossman

Chris

2 years

🎉 New feature alert! 🎉 Say hello to effortless data joining with . Easily add data to your conversations and answer even the most challenging data questions. 💻🔍

Tweet media one

0

0

4

@cdossman

Chris

2 years

Really like the traceback of actions the system is taking to achieve a solution.

@DYtweetshere

DY

2 years

AI agents can now teach themselves HOW to use tools (ie. any API) in real time, completely automated! Introducing: Self-Learning Agent for Performing APIs (SLAPA) with @FinsamSamson

106

356

2K

1

1

6

@cdossman

Chris

2 years

You miss all the shots you don’t take! @agazdecki thanks for bringing us along for the journey.

@agazdecki

Andrew Gazdecki

2 years

Securing @acquire has a chance! Answer is always no unless you ask!

59

15

407

1

0

6

@cdossman

Chris

10 months

@yacineMTB Have had a few bad trips out there, dealt first hand with the crime. Looks like I’ll be spending 2 months out there in June and July. Hopefully I get a better experience, looking forward to meeting people

3

0

6

@cdossman

Chris

5 years

Quantization effects on deep reinforcement learning using Quantized Reinforcement Learning (QUARL). Amazing research work by @Srivatsan91 @sharad24sc , Maximilian Lam, Zishen Wan, @AleksandraFaust , and Vijay Janapa Reddi

Tweet card media

What if Quantization was Applied for Reinforcement Learning?

Reinforcement learning (RL) has reached a place where it can be applied to many real-world applications such as robotics, game playing…

0

4

6

@cdossman

Chris

2 years

Imagine a world where you can ask data questions and get answers based on your companies data. No more tedious work and way faster answers when you need it. This Wednesday the demo will drop.

0

0

6

@cdossman

Chris

11 months

@TheNextBigRush Tell that to the elderly living on fixed incomes who can’t afford the cost of living increase because of all of the tech bros moving to Portugal 🇵🇹

2

0

6

@cdossman

Chris

1 year

See you guys next Sunday

Tweet media one

0

0

6

@cdossman

Chris

2 years

What are some must-follow Twitter accounts in the AI space? I'll go first: 1. @jackclarkSF for AI industry weekly newsletter and related tweets 2. @nathanbenaich for the state of AI and deep tech VC 3. @_akhaliq stream of awesome research papers

1

0

5

@cdossman

Chris

1 year

Who is working on open source decentralized LLM api? 10 million + GPUs in the world that can run 3.5 turbo like models.

3

1

5

@cdossman

Chris

1 year

@demishassabis @ylecun Way before we get to that level of power gen we would start to run in to issues getting rid of the excess heat.

0

0

5

@cdossman

Chris

1 year

Anecdotally, Claude-2 has a hallucination problem I don’t see with GPT4. For coding this makes it pretty hard to rely on its completions even when giving it already correct requests.

0

1

4

@cdossman

Chris

2 years

Was just looking in to this for a content app I’m working with. Excited to try it out. Will share results.

@DynamicWebPaige

👩‍💻 Paige Bailey

@DynamicWebPaige

2 years

Session-based recommenders give recommendations based solely on a user’s interactions in an ongoing session—and do not require the existence of user profiles, or historical preferences. 💓 this short course showing how to build them with @RayDistributed :

Tweet media one

1

23

131

0

1

5

@cdossman

Chris

11 months

GM ☀️ What is everyone building? Today I’m fine tuning clip on 200,000 ad creatives

0

0

5

@cdossman

Chris

1 year

The guy who plagiarized research papers and course material, just launched a medical focused fine tuned Llama2 model. Stay away from the product and grab some popcorn. Should be interesting how this story unfolds.

@sirajraval

Siraj Raval

1 year

Introducing DoctorGPT! After applying fine-tuning, reinforcement learning, & compilation techniques to Meta's Llama2 model, I got amazing results: - Passes the US Medical Licensing Exam - Offline - iOS & Android - Open Source Code: Full video tutorial:

132

518

3K

0

0

5

@cdossman

Chris

5 years

How to Enable Robots Understand Incomplete Natural Language with Common Sense Reasoning. Amazing research work by @mohitban47 @Roni_Aerospace @chenhaonani1995 Alan Kuntz, and Hao Tan @uncnlp

0

2

5

@cdossman

Chris

2 years

🏇 Sunday Launch 🤖 💬 Natural Language Answers without visualizations 🧪 Situational Suggestions More Chart Support Coming Soon - Histogram - Doughnut - Polar Area - Radar - Scatter - Bubble Try the open alpha at ⏩

Tweet media one

1

0

4