Connect with us

Technology

Why multi-agent AI tackles complexities LLMs can’t

Published

on

Why multi-agent AI tackles complexities LLMs can't

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More


The introduction of ChatGPT has brought large language models (LLMs) into widespread use across both tech and non-tech industries. This popularity is primarily due to two factors:

  1. LLMs as a knowledge storehouse: LLMs are trained on a vast amount of internet data and are updated at regular intervals (that is, GPT-3, GPT-3.5, GPT-4, GPT-4o, and others);
  1.  Emergent abilities: As LLMs grow, they display abilities not found in smaller models.

Does this mean we have already reached human-level intelligence, which we call artificial general intelligence (AGI)? Gartner defines AGI as a form of AI that possesses the ability to understand, learn and apply knowledge across a wide range of tasks and domains. The road to AGI is long, with one key hurdle being the auto-regressive nature of LLM training that predicts words based on past sequences. As one of the pioneers in AI research, Yann LeCun points out that LLMs can drift away from accurate responses due to their auto-regressive nature. Consequently, LLMs have several limitations:

  • Limited knowledge: While trained on vast data, LLMs lack up-to-date world knowledge.
  • Limited reasoning: LLMs have limited reasoning capability. As Subbarao Kambhampati points out LLMs are good knowledge retrievers but not good reasoners.
  • No Dynamicity: LLMs are static and unable to access real-time information.

To overcome LLM’s challenges, a more advanced approach is required. This is where agents become crucial.

Agents to the rescue

The concept of intelligent agent in AI has evolved over two decades, with implementations changing over time. Today, agents are discussed in the context of LLMs. Simply put, an agent is like a Swiss Army knife for LLM challenges: It can help us in reasoning, provide means to get up-to-date information from the Internet (solving dynamicity issues with LLM) and can achieve a task autonomously. With LLM as its backbone, an agent formally comprises tools, memory, reasoning (or planning) and action components.

Components of an agent (Image Credit: Lilian Weng)

Components of AI agents

  • Tools enable agents to access external information — whether from the internet, databases, or APIs — allowing them to gather necessary data.
  • Memory can be short or long-term. Agents use scratchpad memory to temporarily hold results from various sources, while chat history is an example of long-term memory.
  • The Reasoner allows agents to think methodically, breaking complex tasks into manageable subtasks for effective processing.
  • Actions: Agents perform actions based on their environment and reasoning, adapting and solving tasks iteratively through feedback. ReAct is one of the common methods for iteratively performing reasoning and action.

What are agents good at?

Agents excel at complex tasks, especially when in a role-playing mode, leveraging the enhanced performance of LLMs. For instance, when writing a blog, one agent may focus on research while another handles writing — each tackling a specific sub-goal. This multi-agent approach applies to numerous real-life problems.

Role-playing helps agents stay focused on specific tasks to achieve larger objectives, reducing hallucinations by clearly defining parts of a prompt — such as role, instruction and context. Since LLM performance depends on well-structured prompts, various frameworks formalize this process. One such framework, CrewAI, provides a structured approach to defining role-playing, as we’ll discuss next.

Advertisement

Multi agents vs single agent

Take the example of retrieval augmented generation (RAG) using a single agent. It’s an effective way to empower LLMs to handle domain-specific queries by leveraging information from indexed documents. However, single-agent RAG comes with its own limitations, such as retrieval performance or document ranking. Multi-agent RAG overcomes these limitations by employing specialized agents for document understanding, retrieval and ranking.

In a multi-agent scenario, agents collaborate in different ways, similar to distributed computing patterns: sequential, centralized, decentralized or shared message pools. Frameworks like CrewAI, Autogen, and langGraph+langChain enable complex problem-solving with multi-agent approaches. In this article, I have used CrewAI as the reference framework to explore autonomous workflow management.

Workflow management: A use case for multi-agent systems

Most industrial processes are about managing workflows, be it loan processing, marketing campaign management or even DevOps. Steps, either sequential or cyclic, are required to achieve a particular goal. In a traditional approach, each step (say, loan application verification) requires a human to perform the tedious and mundane task of manually processing each application and verifying them before moving to the next step.

Each step requires input from an expert in that area. In a multi-agent setup using CrewAI, each step is handled by a crew consisting of multiple agents. For instance, in loan application verification, one agent may verify the user’s identity through background checks on documents like a driving license, while another agent verifies the user’s financial details.

Advertisement

This raises the question: Can a single crew (with multiple agents in sequence or hierarchy) handle all loan processing steps? While possible, it complicates the crew, requiring extensive temporary memory and increasing the risk of goal deviation and hallucination. A more effective approach is to treat each loan processing step as a separate crew, viewing the entire workflow as a graph of crew nodes (using tools like langGraph) operating sequentially or cyclically.

Since LLMs are still in their early stages of intelligence, full workflow management cannot be entirely autonomous. Human-in-the-loop is needed at key stages for end-user verification. For instance, after the crew completes the loan application verification step, human oversight is necessary to validate the results. Over time, as confidence in AI grows, some steps may become fully autonomous. Currently, AI-based workflow management functions in an assistive role, streamlining tedious tasks and reducing overall processing time.

Production challenges

Bringing multi-agent solutions into production can present several challenges.

  • Scale: As the number of agents grows, collaboration and management become challenging. Various frameworks offer scalable solutions — for example, Llamaindex takes event-driven workflow to manage multi-agents at scale.
  • Latency: Agent performance often incurs latency as tasks are executed iteratively, requiring multiple LLM calls. Managed LLMs (like GPT-4o) are slow because of implicit guardrails and network delays. Self-hosted LLMs (with GPU control) come in handy in solving latency issues.
  • Performance and hallucination issues: Due to the probabilistic nature of LLM, agent performance can vary with each execution. Techniques like output templating (for instance, JSON format) and providing ample examples in prompts can help reduce response variability. The problem of hallucination can be further reduced by training agents.

Final thoughts

As Andrew Ng points out, agents are the future of AI and will continue to evolve alongside LLMs. Multi-agent systems will advance in processing multi-modal data (text, images, video, audio) and tackling increasingly complex tasks. While AGI and fully autonomous systems are still on the horizon, multi-agents will bridge the current gap between LLMs and AGI.

Abhishek Gupta is a principal data scientist at Talentica Software.

Advertisement

DataDecisionMakers

Welcome to the VentureBeat community!

DataDecisionMakers is where experts, including the technical people doing data work, can share data-related insights and innovation.

If you want to read about cutting-edge ideas and up-to-date information, best practices, and the future of data and data tech, join us at DataDecisionMakers.

Advertisement

You might even consider contributing an article of your own!

Read More From DataDecisionMakers


Source link
Continue Reading
Advertisement
Click to comment

You must be logged in to post a comment Login

Leave a Reply

Technology

Apple TV Plus reportedly cancels one of its best new sci-fi shows, Sunny – leaving the series on a cliffhanger ending

Published

on

Rashida Jones looks at a robot in Sunny

The future isn’t looking bright for Sunny, as Apple TV Plus has reportedly cancelled the critically acclaimed sci-fi series after one season.

Screen Daily was the first to report the rumor, with the publication writing: “Sources close to the production confirmed that the series, produced by A24, will not return following the 10-episode run that began in July and concluded with a cliffhanger on September 4.” TechRadar has contacted Apple TV Plus for comment, and we’ll update this article if we get a response.

Source link

Advertisement

Continue Reading

Technology

Xbox Game Pass is getting tons of new games following Black Ops 6

Published

on

Xbox Game Pass is getting tons of new games following Black Ops 6

There’s nothing more refreshing than taking a nice big swig of a Perk-A-Cola during an intense Zombies match with your pals in Call of Duty: Black Ops 6. Besides tasting great (we assume), these colas empower you with buffs that will keep you alive until the later rounds. These bonuses are just as important as the Pack-A-Punch, but come in a wider range of flavors. It is possible to stack them all, but that gets expensive very fast so you need to be thrifty about choosing which ones are most valuable early on. These colas come in eight different types that may or may not be worth your Essence to chug. If you want to know which ones are the best in Black Ops 6, check out our ranking.
Best Perk-A-Colas in Black Ops 6

We’re ranking these Perk-A-Colas based on their base bonuses, not any additional effects you can get if you Augment them, from best to worst.
Jugger-Nog
You can never go wrong with the classic Jugger-Nog. Increasing your health by 100 is useful for the entire duration of a Zombies mode, but especially in the early rounds. Going down is at best a major setback and at worst a cascade into defeat. This extra health allows you to get out of a bad situation that would otherwise end your run.
Quick Revive
This Perk has two components, with the less useful one being the ability to revive teammates 50% faster. Ideally, that shouldn’t be needed but is nice. However, what makes it rank so high is the 50% shorter delay on health regen. Your health will come back over time in Zombies but at a dreadful pace. Starting to heal faster means less time running and kiting zombies around and more time in the action helping the team.
Speed Cola
The best early guns for Zombies mode have a large magazine, fast reload, or both. Running out of ammo at the wrong time and not being able to get more rounds in the chamber fast enough has ended many a run, but Speed Cola is the solution. It will speed up your reloads and armor plating by 30%. That’s not a massive number, but fast enough to make the difference.
Stamin-Up
These zombies aren’t the shambling types. They will chase you down like sprinters, and some of the monstrous ones can outpace you even at full tilt. Stamin-Up sounds like it would let you run for longer, but actually just lets you run faster since you have no limit on sprint in Black Ops 6. Positioning is everything in Zombies so being able to get where you need to be faster is never a bad thing, especially if it’s a downed teammate bleeding out.
Deadshot Daiquiri
Every zombie has a weak spot, which is most often the head. Deadshot Daiquiri makes a critical hit hurt even more than normal, plus increases your auto-aim on those weak points when you aim down sights (ADS). That second part shouldn’t factor much if you’ve got good aim, which is why this is only a decent Perk-A-Cola.
PhD Flopper
Not being able to hurt yourself is situationally useful, and the ability to cause an explosion by diving is a risky move. Yes, it’s cool, but you don’t want to be lying prone surrounded by zombies. This is a cool gimmick, but save yourself the Essence and skip it.
Elemental Pop
When Elemental Pop works, it’s great. The problem with it is that you can’t rely on it. It gives a small chance on every shot you take to add a random ammo mod effect, which is cool but impractical. If you have spare Essence on you, sure, it won’t hurt you to have, but it isn’t something that will turn the tide for you or your team.
Melee Macchiato
Finally, Melee Macchiato bottoms out the list. It is fine in the early game when the hordes are small and you’re better off meleeing enemies since it will smack enemies away with each hit, but quickly drops off in usefulness. When the zombies get too strong and too numerous, you will need a lot more than a strong punch to deal with them and end up forgetting you even have it.

Source link

Continue Reading

Technology

Best Apple iPad to Buy in November 2024

Published

on

Best Apple iPad to Buy in November 2024

Apple has really bloated its iPad lineup in recent years. Essentially, there are four different lines of iPads that you can choose from. There’s the iPad, iPad Air, iPad Mini, and iPad Pro. All of these are here for different use cases and different people. The regular iPad is almost solely for those in education (schools, school districts, etc). At the same time, most will buy the iPad Air or Mini, with the iPad Pro being more so for those “pro” users who might be using it to replace their laptop.

This is why we have come up with this buyer’s guide for the iPad. Helping you get the best iPad for your use-case. So you can pick out the best iPad and not waste any money. So here are the best iPads you can buy in 2023.

Best iPad for most people

6533554cv11d

iPad Air (6th Gen)

The iPad Air has finally been updated after two long years. Now, the iPad Air comes in two sizes – 11-inch and 12.9-inch. It also comes with the M2 chipset. So it’s a pretty powerful iPad, making it perfect for most people. The 11-inch iPad Air starts at $599 for 128GB of storage, with the 12.9-inch model starting at $799. Previously, the iPad Air was available in either 64GB or 256GB at these same prices. So definitely a nice upgrade. You can opt for 256GB, 512GB, and 1TB as well.

If you’re thinking about getting the iPad Air versus the iPad Pro, there are a few differences. For instance, the iPad Air does have a fingerprint sensor for Touch ID, instead of using Face ID. It also has a single camera on the rear, with no LiDAR sensor. The only other major difference is going to be the display. The new iPad Pro models have OLED displays, while the iPad Air sticks with LCD. There is a pretty big difference, and if you are planning to use this for a lot of streaming, you might want to splurge on the iPad Pro. Additionally, it sports the M4 processor over the iPad Air’s M2. That won’t translate to much of a difference in terms of performance because iPad OS can’t really take full advantage of the chipset’s power – yet.

Advertisement

Interestingly, now that the iPad Pros are so thin and light, the iPad “Air” is not the thinnest nor lightest iPad that Apple sells. The iPad Pro is both thinner and lighter. So that is something else to consider if you really want to get an iPad that is lightweight.

However, the iPad Air is a really great option for those who want a family iPad. That’s because it’s great for streaming media and also great for getting some work done with the Magic Keyboard, as well as a few other great third-party keyboards from Logitech, among other companies. It won’t replace a laptop, but it can work in a pinch.

Buy at Best Buy

Best Cheap iPad

GUEST 430900c7 e560 448c bc9c 2b605b9f72a1

iPad (10th Gen)

  • Price: $449
  • Released: October 2022
  • Where to buy: Target

The iPad tenth generation is the best cheap iPad that you can buy today. It is the cheapest new iPad available. It’s not the best iPad, as it does have some interesting quirks, like the fact that it has a USB-C port and still supports the original Apple Pencil, which charges via Lightning. So, you need a couple of dongles to use the Apple Pencil with this iPad. But thankfully, Apple has made it a bit easier now, since you can get the Apple Pencil (USB-C), and only need a single cable to sync and charge your Apple Pencil, but you would lose pressure sensitivity.

This iPad is actually pretty good, all things considered. It’s the same size as the iPad Air, coming in at 10.9 inches, though the display isn’t quite as good, which is how Apple was able to keep the price down here. It also has the A14 Bionic chipset, which was in the iPhone 12 series, so it’s a few years older at this point.

Advertisement

Perhaps the most interesting aspect of this iPad is the fact that Apple gave us a horizontal front-facing camera. So it’s not in the right place for Zoom, Google Meet and Facetime. Whereas it is in the wrong place on virtually every other iPad. It does have a 12-megapixel camera on the front and back, so you can use it to take photos, but we would not recommend that.

Interestingly, the iPad 10th-generation does have it’s own magic keyboard folio, which many say is better than the actual Magic Keyboard available for the iPad Pro and Air models. It’s pretty interesting, but it is a good way to get a keyboard for your new iPad. The iPad 10th-generation, much like the iPad Air, only comes in 64GB and 256GB, with a $150 price difference between the two. So while it starts at $449, the one you would probably want, starts at $599.

iPad (10th Generation) – Target

Best iPad for a laptop replacement

7147JzEtrqL AC SL1500

iPad Pro 13

Apple has finally gone OLED with iPads. With the new iPad Pro in 2024, Apple has added OLED to the “Pro” iPads, it’s unclear when or if Apple will OLED to its other iPads, but this addition does make the iPad Pro a bit pricier this time around. So there are two models available now, an 11-inch that starts at $999 and a 13-inch that is $1299. These come with 256GB of storage by default, so much like the iPad Air, the Pro’s also got a bump in its starting storage.

Advertisement

For both of the iPad Pro models, you can get 512GB, 1TB, and 2TB storage options as well. The 1TB and 2TB options include a bump in RAM up to 16GB and an extra processing core on the M4 chipset. Those two higher-end models also include an option for a nanotexture display. This gets rid of reflections quite a bit, but it does degrade the quality of the display a bit. So that’s something to keep in mind.

I personally bought the iPad Pro 13 for myself, and I absolutely love this size for an iPad. It’s great for travel, and in fact, I’ve traveled with it only instead of bringing my MacBook Pro, and it’s been a great experience. That OLED display is insanely bright, especially for OLED. Of course, a bit part of that is because Apple is using Tandem OLED here, essentially, what Apple is doing is using two OLED displays here. This allows the display to get much brighter without destroying the color profile of the OLED. Which is one of the main selling points of OLED. The iPad Pro 13 (and 11-inch) get up to 1,000 nits of peak brightness. That makes it really great to use outdoors.

Buy at Best Buy

61IsQRFoxJL AC SL1500

iPad Pro 11

The new iPad Pro 11 is basically identical to the 13-inch, with the only real change being the fact that it is smaller. Previously, the larger iPad Pro had a different display entirely, but that is no longer true. Both the 11-inch and 13-inch iPad Pros are using Tandem OLED displays. So you can get the same great display on both tablets. And that’s truly incredible.

Advertisement

With Tandem OLED, that is making the OLED display on the iPad Pro brighter, without sacrificing the color. This is because you’re basically getting two OLED panels here. The color reproduction on this panel is really incredible, and if you’re like me and use your iPad for a lot of streaming, then you’ll really notice the difference here.

But, this is going to be hamstrung by iPadOS, which is still pretty lacking. It’s almost like Apple is intentionally not unleashing iPadOS to keep the Macs selling like hot cakes. But if you want something that is close to a computer replacement, the iPad Pro 11 is another great option.

iPad Pro 11 – Amazon

Best iPad for reading and travel

71o6PxcKkJL AC SL1500

iPad Mini (6th Generation)

  • Price: $499
  • Released: September 2021
  • Where to buy: Amazon

The iPad Mini is surprisingly one of Apple’s most popular iPads. It’s surprising given the size, since it does have a 8.3-inch display. Making it only about an inch and a half larger than Apple’s most popular iPhone model – the Pro Max. But many love this iPad Mini since it is great for travel and reading eBooks, as well as watching content while on the go. Pilots absolutely love this iPad, since it is small enough to fit in the cockpit without any issues.

iPad Mini is a bit older at this point, but surprisingly, it has a more recent chipset than the regular iPad that’s only $50 cheaper. The iPad Mini sports the A15 Bionic chipset. This first debuted in the iPhone 13 series back in 2021. It does also have 12-megapixel cameras on the front and back, for taking photos and also using it for video calls on Facetime, Zoom and Google Meet (among others). It does also have Touch ID, much like every other iPad that is not the iPad Pro. Honestly, this is probably the best way for authentication. Since it is in the power button. Something you’re pressing anyways.

Advertisement

This iPad does also support the Apple Pencil 2nd-generation and Apple Pencil (USB-C), it does not support the first-generation Apple Pencil, because it’s Lightning. We’d recommend pairing this with the Apple Pencil second-generation. Apple does not sell any keyboard for the iPad Mini, but it does support Bluetooth keyboards. We wouldn’t recommend buying one, since the iPad Mini is so small, any keyboard case is going to be very cramped. If you do get one, don’t get a keyboard case, but an external keyboard.

Finally, let’s talk pricing. The iPad Mini comes in two storage capacities. That’s 64GB for $499 or 256GB for $649. Apple does also offer 5G connectivity for another $150 (plus what your carrier is going to charge you). Finally, the iPad Mini comes in four colors: Space Gray, Pink, Purple, and Starlight.

iPad Mini (6th Generation) – Amazon

Source link

Advertisement

Continue Reading

Technology

Kindle Paperwhite Signature (2024) review: A luxurious reading experience

Published

on

Kindle Paperwhite Signature (2024) review: A luxurious reading experience

Amazon completely revamped its Kindle family in October when it debuted the entirely new Colorsoft ereader along with updated generations of the other three existing models: the standard Kindle, the Paperwhite and the Scribe E Ink tablet. The new Paperwhite has a slightly larger screen, a bigger battery and a more powerful processor that Amazon combined with E Ink tech to make page turns faster. This Paperwhite again comes in a Signature Edition, which adds wireless charging, an auto-adjusting front light and extra storage — all for $200, which is $10 more than the previous Signature Edition.

The standard Kindle also got a $10 price bump, leaving the same $90 gap between Amazon’s cheapest model and the Signature Edition. I tested both to see whether the price difference between the two is worth it. The higher-end model is certainly posh, but the base model handles ereader basics well. Is the deluxe treatment that much better?

Amazon/Engadget

The Kindle Paperwhite Signature Edition is one of the most luxe ereaders you can buy, but some may not need to pay extra for all the nice-to-have features.

Pros

  • Screen displays crisp text with a lovely warm light
  • Flush-front screen and soft-touch back feel premium
  • Noticeably faster page turns and scrolling
  • Auto-adjusting warm front light is helpful
Cons

  • Power button on the bottom edge is easy to accidentally trigger
  • Can’t adjust touch screen controls
  • Fewer page and font adjustments than the competition
  • The features exclusive to the Signature aren’t essential

$200 at Amazon

Amazon/Engadget

The $110 Kindle is a no-frills, highly portable ereader that covers all the basics well. The lack of a warm front light is its only major flaw.

Advertisement

Pros

  • The most affordable ereader
  • Compact and durable build is highly portable
  • Grants access to Kindle’s vast ebook catalog plus library books
Cons

  • No warm light
  • Not waterproof
  • Can’t read third party ebooks

$110 at Amazon

The company is calling the Paperwhite (Signature and standard) the “fastest Kindle ever” thanks to a new dual-core processor and an oxide thin-film transistor layer that enables quicker switching of the text and graphics on the page. That layer is also more transparent than the previous tech, so the on-page clarity has increased as well.

As for storage, the Paperwhite Signature gets 32GB, double that of the regular Paperwhite, which is the same treatment as the prior generations for both. They have a couple additional backlights this time: ten standard LEDs and nine warm ones for a total of 19.

The screen is a little larger at seven inches, up from 6.8, and sports the same 300 ppi resolution. Despite a larger battery capacity (with an estimated 12 weeks of live instead of just ten), the device is a fraction of a millimeter thinner than last time, and weighs just two-tenths of a gram more. It has the same IPX8 rating, meaning it can handle full submersion, and the back is made from a metallic plastic with a rubberized texture that’s easy to grip (so you’re less likely to accidentally dunk it in water). The bezels help with grip too, as they’re just wide enough to accommodate a thumb without triggering a page turn.

The hardware upgrades are subtle, but they add up to a downright elegant way to read books. Pulling up menus, opening new titles and turning pages are all speedier than any ereader I’ve tried. And ghosting is basically non-existent. The roomy screen displays text that’s crisper than a January morning and the auto-adjusting front lights are like a warm lantern lighting your way.

If I take any issue with the design of the Signature, it’s the placement of the power button at the bottom edge. One of the simple joys in this life (for me anyway) is eating while reading. That means I hold a book or ereader on the table in front of my plate as I shovel stir fry into my mouth. Often, resting the Signature on the bottom edge accidentally presses the power button, shutting off my read mid-forkful. Notably, this doesn’t happen with the standard Kindle, even though the button is in the same place, simply because the smaller ereader isn’t heavy enough to trigger the button-press. The Signature also has a rubber grommet around the edge of the flush-front screen, which I don’t hate, but I do find myself distractedly fidgeting with it when I read.

Advertisement

A close up of the text on the Kindle Paperwhite Signature Edition

Photo by Amy Skorheim / Engadget

I sat the Paperwhite Signature next to other ereaders in my testing collection and compared the page-turn speeds, menu load times and scrolling refresh rates. Though the differences weren’t significant, the Signature was definitely faster. (But the Kobo Libra Colour is still the fastest at waking up after a sleep of longer than a few minutes.)

The Signature’s touch response is impeccable. I’ve accepted the fact that most ereader screens occasionally need an extra tap before they do what I want. Whether that’s turning a page or pulling up a menu, my first gesture sometimes goes unnoticed. Not here; the screen immediately responds to every swipe and tap I make. Unfortunately, sometimes that tap is unintentional, and other times the screen doesn’t respond the way I want — but it always responds.

That leads me to a couple of disappointments in the Paperwhite’s interface. Those misinterpreted taps were almost always page turns. The area that skips back in your reading is a narrow sliver on the left-hand side. Some people may plow forward in a book with no backwards glances, but I often find myself at the top of a new page with the realization that I’d been thinking about toast or Cristin Milioti for most of the last page and have no idea what’s going on. Then, when I tap to go back, I’m often taken forward, which confuses me more. To be fair, swipes are generally accurate — but I’m a tapper not a swiper.

This wouldn’t be a problem if the area for a back tap were wider — but you can’t adjust that. In fact, you can’t adjust the way your taps and swipes work at all. The middle of the top of the page accesses the header menu and a swipe down from the top brings up the quick settings panel. There’s no switching those gestures or the zones to suit your preference.

Advertisement

It’s unfortunate because that’s something both Kobo and Boox devices can do. Those readers also give you finer adjustments for the margins, line spacing, font weight and more. The options the Kindle does provide for the look of your text are honestly fine, though. And the Kindle gets points for letting you create and save themes made up of different combinations of fonts and layouts. Kobo doesn’t allow this and Boox can’t do it in its native reader app.

The Kindle Paperwhite Signature Edition is propped up on a bench near a backpack

Photo by Amy Skorheim / Engadget

Acquiring books you want to read is maybe one of the most important qualities in an ereader and Kindles have their strengths and weaknesses. These features aren’t unique to the Paperwhite, but it’s worth mentioning how Amazon’s book access compares to its competitors. The first thing to acknowledge is that the Amazon Kindle library is the largest, thanks to Amazon Exclusive books and self-published, Kindle Direct Publishing (KDP) titles. That said, if you mostly read titles from established publishers, you’ll find them on any reader.

Amazon’s devices are better for people who like to flit between reading and listening to a given book. Not only do you get a discount on an audiobook if you buy the ebook version, the Whispersync feature keeps track of where you are so you can swap between the formats and pick up the story at the right place. And if you’re active on Goodreads, a Kindle will also serve you better as the integration with (the Amazon-owned) site is baked-in.

What Kindles don’t support are ePubs with any digital rights management (DRM) other than its own. If you buy a book from a third party ebook store that uses Adobe DRM, you can’t read it on a Kindle — but you can on Kobo and Boox devices (with a few extra steps). Finally, you may want to consider how you’ll access books borrowed from your local library. Using either the Libby app on your phone or your local library’s website, you just need to choose the “send to Kindle” option for them to show up on your reader. It’s worth noting that Kobo lets you search for and borrow library books from the device itself, and Boox devices let you do so from the Libby app directly.

Advertisement

The Kindle Paperwhite Signature Edition is propped up next to the regular kindle

Photo by Amy Skorheim / Engadget

The regular Paperwhite is $160 and the Signature Edition is $200. That’s a $40 increase that may make sense for some, but probably not for most people. The three add-ons (extra storage, wireless charging and auto-adjusting light) are handy, but won’t dramatically improve your reading experience. If you like having offline access to lots of audiobooks, the larger 32GB capacity might be worth it. But if you mostly read ebooks (and/or use your phone for audiobooks, which makes more sense to me), the 16GB-capacity should be more than enough for years of library expansion.

The auto-adjusting front light on the Signature Edition is responsive and calibrates the light to the environment pretty well. I still need to manually lower it in a fully dark room, so it’s not fully hands-off, plus it’s easy enough to adjust. The separate warm light (which you get with the regular Paperwhite) is the more critical feature anyway, as it makes reading into the wee hours of the night far more pleasant. The final Signature-only feature, wireless charging, is probably the least important — you only charge these things once every other month at the most. How you do so feels inconsequential.

So if you save $40 by not going with the Signature model, does it make sense to save even more by going with the $110 standard Kindle? That’s harder to answer. The base Kindle has the same 300 ppi resolution, the same storage capacity, the same font and layout options and access to the same audio- and ebooks as the standard Paperwhite. But the Kindle’s screen is smaller, the device is not waterproof, the battery lasts half as long and there’s no warm light. Touch responsiveness isn’t quite as good and page turns are slightly slower (though the difference on both counts is minor).

Honestly, the lack of a warm light on the base Kindle is the only major drawback. It’s far more enjoyable to read the softer yellowish screens of either Paperwhite than the harsher blue of the base model. But I found myself getting used to it fairly quickly, and I probably noticed the difference most when switching back and forth between the two — something most people won’t do.

Advertisement

The new Kindle rests on a backdrop of moss and stone

Photo by Amy Skorheim / Engadget

In short, the Kindle Paperwhite and the new Paperwhite Signature offer a more premium reading experience. The flush front screen and soft-touch back feel more upscale. The larger screen feels like the difference between reading a mass market paperback and a trade paperback, which isn’t to say the standard Kindle feels cramped, it’s just smaller. I see the Paperwhite (especially the Signature model) appealing to people who want the most elevated reading experience — the crispest text, the least lag, the swankiest feel. And the regular Kindle is perfect for those who just want the most direct line to reading nearly any book they want.

It comes down to whether you think a heightened experience will make you read more. If so, then it’s probably worth the price jump in the long run. But if you’re just looking for a no-frills way to read books, go for the base Kindle. Even though it’s less high-end, it feels more durable and more portable. The smaller size can fit in a back pocket and you may be more apt to toss it in a cluttered bag — all of which may very well mean you read more, and, really, that’s the whole point of any ereader.

Source link

Continue Reading

Technology

Buddy raises $1.1M for social utility app

Published

on

Buddy is a new social utility app.

Buddy is a new social utility app.


Buddy, a social utility app designed to get you off your phone, announced a $1.1 million pre-seed funding round.Read More

Source link

Continue Reading

Technology

No, startups shouldn’t always take the highest valuation, seed VCs say

Published

on

Corinne Riley, Elizabeth Yin

One of the lessons that the wild Silicon Valley venture funding environment of the past few years has clearly taught is this: Bigger valuations are not always better.

“I think we’ve all kind of seen the negative impact of having a valuation too high from the last, call it, three years,” Elizabeth Yin, co-founder of Hustle Fund said onstage at TechCrunch Disrupt last week. When a VC bull market hits and startups are able to easily raise a lot of money before they have real, provable businesses, they’ve set themselves up for difficult times.

Because “the bar is higher for that next round,” she said. The general rule is for each early round, business growth should justify double, or possibly triple, the previous valuation, Yin said.  

So early valuations “shouldn’t be anything really crazy that you don’t think you can grow into realistically with your traction, because it always catches up with you,” she said.

Advertisement

If the company fails to grow into a lofty valuation, it could wind up burning its most valuable employees, said VC Renata Quintini, co-founder of Renegade Partners.

Most startups grant stock to employees, or sometimes grant stock options — which the employee is required to buy. And most startups offer that stock as a significant portion of their employees’ salaries. People join startups because they believe if they help build the company, their stock will pay off. So, obviously, it is not good if employees’ stock grows less valuable over time.

“If that gap doesn’t close, you’re actually disincentivizing the people that joined you early on,” Quintini warns.

A much better way to raise money is to “create a tight process,” by setting reasonable valuation expectations from the get-go, VC Corinne Riley, partner at Greylock, said onstage. “You don’t want to be dillydallying and have a multi-month round. You’re wasting your own time. You’re wasting the VC time,” she said. “You want to know exactly how much you want to raise.” 

Advertisement

Quintini advises founders to have ranges in mind for both an amount and a valuation. To do that, she says, a founder should spend more time in an information-gathering phase than in an actual pitching phase.

They should ask VCs in their network their opinions on their valuation. They should know what type of market they are in and what the multiples on revenue or other pricing metrics are in vogue for their area at the moment. They should carefully consider how much dilution they are willing to take — that is, how much of their company they are willing to sell off and how much of a stake they will retain after the round.

Should the founder want to sell a smaller stake — 10% versus the more typical 20% — the founder should find out which firms would even entertain that idea. Many firms won’t bother with small stakes, as that decreases their chances for a big return. 

Coming into the pitch meeting wanting too much for too little means “you better have a fantastic business and be an outlier company to back it up; otherwise, you’re actually going to be turning VCs off,” Quintini says.

Advertisement

Renata Quintini, Corinne Riley, Elizabeth Yin
Renata Quintini, Corinne Riley, Elizabeth Yin (left to right). Image Credits:Barak Shrama/ Slava Blazer Photography / Flickr (opens in a new window)

If a VC comes in with a term sheet that wildly beats all the others in valuation, founders should look at the fine print. Has the VC banked the term sheet toward giving its firm outsized power? This could also mean that the company won’t be able to convince other VCs to invest in later rounds.

Startup accelerator Y Combinator distributes a sample term sheet that shows off what most VCs consider standard terms. These cover everything from voting rights to board seats.

“I’ve definitely seen a number of my founders, especially international companies, get all kinds of term sheets with all kinds of terms that I would consider nonstandard,” Yin described, such as “weird board configurations” like the VC wanting multiple board seats, or “all kinds of liquidation” preferences. Anything above a “1x” liquidation preference means that the investor gets paid out more money, and first, should the company sell and isn’t standard.

In addition to being prepared to negotiate on dollar amount, valuation, and stake size, founders should be prepared to negotiate board composition and items like who gets to choose the independent board members. Whatever you decide on the terms that give VCs power could impact your company, and its future valuations, forever.

“I encourage our founders to turn the very nonstandard things down. And then there are some others that are borderline. And maybe you take it because you don’t have any other options, but, once it’s done, it’s really hard to unwind,” Yin says.

Advertisement

Source link

Continue Reading

Trending

Copyright © 2024 WordupNews.com