Alexander Hedlund @valueinnovate - Twitter Profile

8 months ago

@EllevioAB Fantastiskt! 40 minuter efter att strömavbrottet började. Är det den KPI ni har på er kontinuitetshantering? Och för oss som inte har egna backupbatterier.

1

0

377

Alexander Hedlund @valueinnovate

8 months ago

@EllevioAB @danjan82 @HenrikIsak1 Hur kommunicerar ni vid strömavbrott?

1

0

42

Alexander Hedlund @valueinnovate

8 months ago

@EllevioAB @danjan82 @HenrikIsak1 När kommer strömmen tillbaka?

0

23

Alexander Hedlund @valueinnovate

8 months ago

@SthlmExergi När kommer strömmen tillbaka?

0

11

Who to follow

Eric Garza

@egarzas

R&D Product Manager at AMI Automation

Shashank Kirtania

@5hv5hvnk

prev @ Microsoft Research, @prosemsft I like cats :3

David Svendsen

@DavidSvendsen

Dad, Veteran, MBA, AI and Vibe Coding Enthusiast, Carnivore. IT Director. https://t.co/VL7zFafPUh

Alexander Hedlund @valueinnovate

8 months ago

strömavbrott stockholm innerstad Vem vet mer?

1

4

1

0

1K

Alexander Hedlund @valueinnovate

over 1 year ago

@AndrewPatriot_ @POTUS When FEAR is part of your governance There is STUPIDITY

0

18

Alexander Hedlund @valueinnovate

over 1 year ago

Great #AiUseCase for #GenerativeAi using in combination with #KnowledgeGraphs for legacy software modernization.

Martin Fowler

@martinfowler

over 1 year ago

FINAL § Alessio, Tom, and Shodhan complete their article on what they have learned from using GenAI with legacy systems. GenAI helps them build a capability map, identifies dead code, and has promise for better translations between platforms. https://t.co/F5e73Rposr

0

169

43

128

18K

0

71

Alexander Hedlund @valueinnovate

over 2 years ago

#AutonomousAi autonomous refinement when needed using an embedded Ai decision pipeline. Read whole thread!

Carlos E. Perez

@IntuitMachine

over 2 years ago

1/n Scientists Invent an "Aha Moment" Generator for AI Systems It's a familiar scene - you've written an essay or solved a math problem, feeling assured you have the right solution. But upon review, it becomes clear your original logic was flawed in places. If only you had the ability to closely re-examine your work and pinpoint exactly where your reasoning diverged from the ideal path. This capacity for self-reflection and targeted improvement remains elusive for even the most advanced AI systems today. In a thought-provoking new paper, researchers propose a method that brings this capability within reach for large language models. Introducing GLoRe - short for Global and Local Refinements - the authors provide a blueprint for how to train AIs to decide when, where and how to refine their own reasoning without any external feedback or supervision. GLoRe broke down a complex cognitive challenge into manageable parts. By strategically combining global reasoning restarts with focused local tweaks, the approach boosted the accuracy of a leading language model by over 10% on complex math problems. Beyond the raw performance gains, GLoRe exhibits something arguably more profound - the inklings of an inner sense for identifying flaws in its own thinking. How is this possible without human oversight? The key innovations involved specialized self-evaluating models and smart pairing of synthetic training data. Together, these unlocked the model's latent capacity for targeted self-improvement. GLoRe demonstrates that the tools for transformational self-reflective learning may already exist within our most capable models today. With the right frameworks, future AIs could refine entire reasoning processes, not just final outputs. The greater promise is of systems that emulate and surpass the human ability to re-examine beliefs, critique arguments, and achieve enlightenment through self-driven growth. If you find that prospect as fascinating as I do, read on to learn how GLoRe represents an important step in this direction.

IntuitMachine's tweet photo. 1/n Scientists Invent an "Aha Moment" Generator for AI Systems

It's a familiar scene - you've written an essay or solved a math problem, feeling assured you have the right solution. But upon review, it becomes clear your original logic was flawed in places. If only you had the ability to closely re-examine your work and pinpoint exactly where your reasoning diverged from the ideal path. This capacity for self-reflection and targeted improvement remains elusive for even the most advanced AI systems today.

In a thought-provoking new paper, researchers propose a method that brings this capability within reach for large language models. Introducing GLoRe - short for Global and Local Refinements - the authors provide a blueprint for how to train AIs to decide when, where and how to refine their own reasoning without any external feedback or supervision.

GLoRe broke down a complex cognitive challenge into manageable parts. By strategically combining global reasoning restarts with focused local tweaks, the approach boosted the accuracy of a leading language model by over 10% on complex math problems. Beyond the raw performance gains, GLoRe exhibits something arguably more profound - the inklings of an inner sense for identifying flaws in its own thinking.

How is this possible without human oversight? The key innovations involved specialized self-evaluating models and smart pairing of synthetic training data. Together, these unlocked the model's latent capacity for targeted self-improvement.

GLoRe demonstrates that the tools for transformational self-reflective learning may already exist within our most capable models today. With the right frameworks, future AIs could refine entire reasoning processes, not just final outputs. The greater promise is of systems that emulate and surpass the human ability to re-examine beliefs, critique arguments, and achieve enlightenment through self-driven growth. If you find that prospect as fascinating as I do, read on to learn how GLoRe represents an important step in this direction.

4

389

97

334

49K

0

45

Alexander Hedlund @valueinnovate

over 2 years ago

Great lessons learned to get more out of your reading!

Alex & Books 📚

@AlexAndBooks_

over 2 years ago

Author @shaneparrish reads 100+ books a year and has a library of 4,000+ books. Here are 37 tips from him on how to read more & better: 1) Reading is a long-term investment I make in myself. I don’t have the time not to. 2) When it comes to reading, my philosophy is simple: Skim a lot of books. Read a few. Immediately re-read the best ones twice. 3) The person who reads one great book twice is better off than the person who reads ten average books. 4) There are no rules when it comes to choosing books. We don’t have to read bestsellers, or classics, or books everyone else raves about. 5) When it comes to reading, you don’t need to finish what you start. As a general rule, people who love reading never, ever finish a crappy book. 6) Speed-reading is bullshit. Getting the rough gist and absorbing the lessons are two different things. Confuse them at your peril. 7) If you’re reading fast you’re not thinking and challenging what you’re reading. You’re not being critical. All of that is work. And if you’re not doing the work, you’re only walking away with surface knowledge. 8) Reading fast is worse than not reading. Reading fast gives you two things that should never mix: surface knowledge and overconfidence. 9) The secret to reading better is reading lots of good books slowly. This enables you to build knowledge. Knowledge, in turn, allows you to read faster with true comprehension and retention. 10) Finding time to read is easier than you might think. Waiting for a bus? Stop staring down the street and read. Waiting for a taxi? Read. On the train? Read. On the plane? Read. Waiting for your flight? Read. 11) There’s an advantage to be gained from reading things other people are not reading because you will gain knowledge and insights that not everyone else has. 12) The more interesting and relevant we find a book, the more likely we are to remember its contents in the future. 13) While rereading can seem like a waste of time because there are so many other books to read, this is a misunderstanding of the learning process. The goal is not to read as many books as possible. The goal is to gain as much wisdom as you can. 14) Teaching others is a powerful way to embed information in your mind. Upon completing a book, grab the nearest (willing) person and tell them about what you have learned. 15) If we read but don’t reflect, we only have the illusion of knowledge. Making the knowledge or own means digesting, synthesizing, and organizing the information. 16) Quality matters more than quantity. If you read one book a month but fully appreciate and absorb it, you’ll be better off than someone who skims half the library without paying attention. 17) Book summary services miss the point. A lot of companies charge ridiculous prices for access to vague summaries bearing only the faintest resemblance to anything in the book. Summaries can be a useful jumping-off point to explore your curiosity, but you cannot learn from them the way you can from the original text. 18) Fancy apps and tools are not needed. A notebook, index cards, and a pen will do just fine. 19) We shouldn’t read stuff we find boring. Life is far too short. 20) The work of reading: Read a book while highlighting heavily. Distill highlights into core insights, written in your own words. Put into a searchable format. Review regularly. 21) Ignorance is more expensive than a book. 22) Understanding more of what you read is better than reading faster. 23) Most people read the same new books that everyone else has read, not necessarily for the ideas but for the social reward of being able to talk about them with others. 24) Reading the same thing as everyone else is only going to put the same ideas in your head that everyone else has. 25) Reading speed is a vanity metric. In the real world, no one cares how fast you read or how many books you read last year or last week. All that matters is what you absorb and apply. 26) A good book, like a good wine, deserves to be savored. Find something worth reading, then chew on the ideas slowly and deeply. Simple but not easy. 27) Reading for curiosity is better than reading for self-improvement. 28) If you want new ideas, read old books. 29) All the time you spend reading something bad comes at the expense of reading something good. 30) Reading consumes time. And if we equate time with money, it should not be wasted on bad books. 31) Good writing reads itself you can’t stop reading. Bad writing, on the other hand, feels like a chore. 32) Not everything needs to be read with the same focus and intensity. Some books deserve a skim, while others deserve your complete attention. 33) Read old books. Read the best ones twice. 34) Writing about what you read is the key to turning the experience of reading into knowledge you can use. Writing is reflection. 35) Reading habits don’t need to be complicated; you can start a simple 25 page-a-day habit right now. While it seems small, the gains add up quickly. 36) Instead of the news, we should read good books. More than just read them, we should re-read them. 37) If you’re struggling to stay focused on a particularly difficult or lengthy book, decide to read a mere 25 pages of it a day. It takes only a few minutes to nibble away at a challenging text. Completing a long book in this manner might take months, but at least you will have read it without getting overwhelmed or bored.

AlexAndBooks_'s tweet photo. Author @shaneparrish reads 100+ books a year and has a library of 4,000+ books.

Here are 37 tips from him on how to read more & better:

1) Reading is a long-term investment I make in myself. I don’t have the time not to.

2) When it comes to reading, my philosophy is simple: Skim a lot of books. Read a few. Immediately re-read the best ones twice.

3) The person who reads one great book twice is better off than the person who reads ten average books.

4) There are no rules when it comes to choosing books. We don’t have to read bestsellers, or classics, or books everyone else raves about.

5) When it comes to reading, you don’t need to finish what you start. As a general rule, people who love reading never, ever finish a crappy book.

6) Speed-reading is bullshit. Getting the rough gist and absorbing the lessons are two different things. Confuse them at your peril.

7) If you’re reading fast you’re not thinking and challenging what you’re reading. You’re not being critical. All of that is work. And if you’re not doing the work, you’re only walking away with surface knowledge.

8) Reading fast is worse than not reading. Reading fast gives you two things that should never mix: surface knowledge and overconfidence.

9) The secret to reading better is reading lots of good books slowly. This enables you to build knowledge. Knowledge, in turn, allows you to read faster with true comprehension and retention.

10) Finding time to read is easier than you might think. Waiting for a bus? Stop staring down the street and read. Waiting for a taxi? Read. On the train? Read. On the plane? Read. Waiting for your flight? Read.

11) There’s an advantage to be gained from reading things other people are not reading because you will gain knowledge and insights that not everyone else has.

12) The more interesting and relevant we find a book, the more likely we are to remember its contents in the future.

13) While rereading can seem like a waste of time because there are so many other books to read, this is a misunderstanding of the learning process. The goal is not to read as many books as possible. The goal is to gain as much wisdom as you can.

14) Teaching others is a powerful way to embed information in your mind. Upon completing a book, grab the nearest (willing) person and tell them about what you have learned.

15) If we read but don’t reflect, we only have the illusion of knowledge. Making the knowledge or own means digesting, synthesizing, and organizing the information.

16) Quality matters more than quantity. If you read one book a month but fully appreciate and absorb it, you’ll be better off than someone who skims half the library without paying attention.

17) Book summary services miss the point. A lot of companies charge ridiculous prices for access to vague summaries bearing only the faintest resemblance to anything in the book. Summaries can be a useful jumping-off point to explore your curiosity, but you cannot learn from them the way you can from the original text.

18) Fancy apps and tools are not needed. A notebook, index cards, and a pen will do just fine.

19) We shouldn’t read stuff we find boring. Life is far too short.

20) The work of reading: Read a book while highlighting heavily. Distill highlights into core insights, written in your own words. Put into a searchable format. Review regularly.

21) Ignorance is more expensive than a book.

22) Understanding more of what you read is better than reading faster.

23) Most people read the same new books that everyone else has read, not necessarily for the ideas but for the social reward of being able to talk about them with others.

24) Reading the same thing as everyone else is only going to put the same ideas in your head that everyone else has.

25) Reading speed is a vanity metric. In the real world, no one cares how fast you read or how many books you read last year or last week. All that matters is what you absorb and apply.

26) A good book, like a good wine, deserves to be savored. Find something worth reading, then chew on the ideas slowly and deeply. Simple but not easy.

27) Reading for curiosity is better than reading for self-improvement.

28) If you want new ideas, read old books.

29) All the time you spend reading something bad comes at the expense of reading something good.

30) Reading consumes time. And if we equate time with money, it should not be wasted on bad books.

31) Good writing reads itself you can’t stop reading. Bad writing, on the other hand, feels like a chore.

32) Not everything needs to be read with the same focus and intensity. Some books deserve a skim, while others deserve your complete attention.

33) Read old books. Read the best ones twice.

34) Writing about what you read is the key to turning the experience of reading into knowledge you can use. Writing is reflection.

35) Reading habits don’t need to be complicated; you can start a simple 25 page-a-day habit right now. While it seems small, the gains add up quickly.

36) Instead of the news, we should read good books. More than just read them, we should re-read them.

37) If you’re struggling to stay focused on a particularly difficult or lengthy book, decide to read a mere 25 pages of it a day. It takes only a few minutes to nibble away at a challenging text. Completing a long book in this manner might take months, but at least you will have read it without getting overwhelmed or bored.

27

521

120

724

155K

0

1

0

98

Alexander Hedlund @valueinnovate

over 2 years ago

Great #ChatGPT #prompt thread

Carlos E. Perez

@IntuitMachine

over 2 years ago

26 Prompting Tips 1 - No need to be polite with LLM so there is no need to add phrases like “please”, “if you don’t mind”, “thank you”, “I would like to”, etc., and get straight to the point. 2 - Integrate the intended audience in the prompt, e.g., the audience is an expert in the field. 3 - Break down complex tasks into a sequence of simpler prompts in an interactive conversation. 4 - Employ affirmative directives such as ‘do,’ while steering clear of negative language like ‘don’t’. 5 - When you need clarity or a deeper understanding of a topic, idea, or any piece of information, utilize the following prompts: o Explain [insert specific topic] in simple terms. o Explain to me like I’m 11 years old. o Explain to me as if I’m a beginner in [field]. o Write the [essay/text/paragraph] using simple English like you’re explaining something to a 5-year-old. 6 - Add “I’m going to tip $xxx for a better solution!” 7 - Implement example-driven prompting (Use few-shot prompting). 8 - When formatting your prompt, start with ‘###Instruction###’, followed by either ‘###Example###’ or ‘###Question###’ if relevant. Subsequently, present your content. Use one or more line breaks to separate instructions, examples, questions, context, and input data. 9 - Incorporate the following phrases: “Your task is” and “You MUST”. 10 - Incorporate the following phrases: “You will be penalized”. 11 - Use the phrase ”Answer a question given in a natural, human-like manner” in your prompts. 12 - Use leading words like writing “think step by step”. 13 - Add to your prompt the following phrase “Ensure that your answer is unbiased and does not rely on stereotypes”.

31

855

144

2K

326K

0

33

Alexander Hedlund @valueinnovate

over 2 years ago

Important Strategic Risk mitigation by @satyanadella @Microsoft A lot of @Azure #Ai services are built on @OpenAI services. So a continuity / contingency approach was a necessity!

Satya Nadella

@satyanadella

over 2 years ago

We remain committed to our partnership with OpenAI and have confidence in our product roadmap, our ability to continue to innovate with everything we announced at Microsoft Ignite, and in continuing to support our customers and partners. We look forward to getting to know Emmett Shear and OAI's new leadership team and working with them. And we’re extremely excited to share the news that Sam Altman and Greg Brockman, together with colleagues, will be joining Microsoft to lead a new advanced AI research team. We look forward to moving quickly to provide them with the resources needed for their success.

4K

87K

14K

5K

42M

0

24

valueinnovate retweeted

Satya Nadella

@satyanadella

over 2 years ago

We remain committed to our partnership with OpenAI and have confidence in our product roadmap, our ability to continue to innovate with everything we announced at Microsoft Ignite, and in continuing to support our customers and partners. We look forward to getting to know Emmett Shear and OAI's new leadership team and working with them. And we’re extremely excited to share the news that Sam Altman and Greg Brockman, together with colleagues, will be joining Microsoft to lead a new advanced AI research team. We look forward to moving quickly to provide them with the resources needed for their success.

4K

87K

14K

5K

42M